Tunnel Sci-Fi Trilogie (Eftos-Epos) II Halbwelt (Adult-Gesamt)
II een fi YES 0 NO El 3 I 0 IC CI AUTHOR SOURCE CODE ...Mar 69 II een fi 3 rbil 1 E. N.) fi \ii...
Transcript of II een fi YES 0 NO El 3 I 0 IC CI AUTHOR SOURCE CODE ...Mar 69 II een fi 3 rbil 1 E. N.) fi \ii...
OE FORM 6000, 2/69 DEPARTMENT OF HEALTH, EDUCATION, AND WELFAREOFFICE OF EDUCATION
ERIC ACC. NO.ED 032 291
mrtn,.. rumr-vrc I nr.=IJIVIC.
IS DOCUMENT COPYRIGHTED? YES 0 NO 1:1.. FO REPRODUCTION RELEASE? YES 0 NO El(IL 61 EL OF AVAILABILITYL I TY I 0 IC "CI
CH ACC. NO.
SP 003 095
P.A.
24
PUBL. DATE
Mar 69
II een fi 3
rbil 1 E. N.) fi \ii
AUTHOR
McDonald, Frederick J.; Koran, Mary Lou.TITLE
The Effect of Individual. Differences on Observational Learning in theAcquisition of a Teaching Skill. Final Report.
SOURCE CODE INSTITUTION ( BOURG E)CI02500 Stanford Univ., Calif., School of Education
SP. AG. CODE SPONSORING AGENCY
MO 004 Office of Education (DREW), Washington, D.C., Bureau of Research'CDR S PRICE CONTRACT NO. G$ANT NO.1.00;11.15
OEG- -8-0810 -011 -010REPORT NO.BUR EAU NO.
BR-8-1-073AVAILABILITY
JOURNAL CITATION
DESCRIPTIVE NOTE
221p..
.
DESCRIPTORS.
.
*Academic Aptitude; *Teaching Skills;.*Training Techniques; Education Majors;Films; Individual Differences; Learning; Microteaching; Teacher Characteristics;Teaching Models
IDENTIFIERS
. .
ABSTRACTA study examined the effects of verbal and perceptual dimensions of individualdifferences in relation to the efficacy of two different kinds of modelingprocedures in the acquisition of a teaching skill (analytic questioning).Aptitude tests for cognitive factors plus specially developed audiovisual testswere administered to 121 intern teachers randomly assigned to three treatmentgroups: a film-mediated modeling treatment (a filmed portrayal of analyticquestioning); a written modeling treatment (a text ofthe film sound track); and acontrol treatment which received no model, but went through all other stepsilcluding initial instructions and microteaching pretest and two cycles of models,rehearsal, and microteaching. The criterion performances assessed by trainedraters included the frequency, variety, and quality of analytic questions used inthree separate teaching sessions in addition to scores on two written posttests.Instructional treatment main effects as well as aptitude by treatment interactionswere investigated using analysis of variance and comparison of regression slopes.Findings, which supported hypotheses, suggest that the rate and level of learningof a specific teaching strategy varies as a function of model presentationfilm-mediated modeling most effective: no modeling least effective); and that the,effectiveness of instructional methods varies from S to S with such differencesbeing related to trainee aptitudes. (ED 017 985 and ED 028 982 are relateddocuments.) (JS)
0I0 O e7oa
U.S. DEPARTMENT OF HEALTH, EDUCATION & WELFARE
OFFICE OF EDUCATION
r\JTHIS DOCUMENT HAS BEEN REPRODUCED EXACTLY AS RECEIVED FROM THE
tr'N PERSON OR ORGANIZATION ORIGINATING IT. POINTS OF VIEW OR OPINIONS
CD STATED DO NOT NECESSARILY REPRESENT OFFICIAL OFFICE OF EDUCATION
POSITION OR POLICY.
Li/FINAL REPORT
Project No. 81073Grant No. OEG-9-8-081073-0117(010)
THE EFFECTS OF INDIVIDUAL DIFFERENCES ON
OBSERVATIONAL LEARNING IN THE
ACQUISITION OF A TEACHING SKILL
Frederick J. McDonald and Mary Lou Koran
School of Education
Stanford University
Stanford, California
March 1969
The research reported herein was performed pursuant to agrant with the Office of Education, U.S. Department ofHealth, Education, and Welfare. Contractors undertakingsuch projects under Government sponsorship are encouragedto express freely their professional judgment in the con-duct of the project. Points of view or opinions stated donot, therefore, necessarily represent official Office ofEducation position or policy.
(09 U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE
(74..Office of Education
ei) Bureau of Research
O0
Qs_
ACKNOWLEDGMENTS
We would like to express our appreciation to
the many people whose help facilitated this study.
We are especially indebted to Dr. Richard E.
Snow, Dr. John D. Krumboltz and Dr. Nathan Maccoby, who
gave generously of their time, effort and encouragement.
Each of these persons, by his effective contribution made
possible the successful conduct of this research. To
them we extend our gratitude even if it can only be ex-
pressed in so modest a form as this acknowledgment.
Frederick J. McDonald
Mary Lou Koran
ii
...
TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS ii
LIST OF TABLES vi
LIST OF FIGURES
SUMMARY
Chapter
I. THE PROBLEM
ix
xi
Definition of the Problem 1
Related Theory and Research 4
Observational Learning 4
Learning and Individual Differences . 9
Implications of Research onObservational Learning andIndividual Differences
Definition of Relevant AbilitiesAbility MeasuresThe Dependent VariableStatement of Hypotheses
1417222730
II. EXPERIMENTAL DESIGN 33
Subjects 33
Ability Measurement 34
Treatment, Procedures 38
Procedures for Model Preparation 42
Lesson Material 43Procedures for Training Microstudents 45
Criterion Measures 47Classroom Performance Measures 47Written Measures 51
Methods of Gathering Data 52Procedures for Training Raters 53
Rating Procedures 54
Rater Reliability 54
iii
7.,74T.MrMri,,,,V7M17,,,MT1,4711+71T,..7.,-","
Page
III. RESULTS 57
Instructional Treatment Main Effects 59Classroom Performance Measures 59Written Measures 79
Ability Measures 82Reliability 82Intercorrelations of Ability Measures . 84Relation of Ability Measures to
Performance 87Aptitude x Treatment Interactions . 92
Classroom Performance Measures 94Written Measures 105
IV. DISCUSSION OF RESULTS AND IMPLICATIONSFOR FUTURE RESEARCH 113
Summary of Data: Hypotheses Tests . . . 1.13
Instructional Treatment Main Effects 114Aptitude x Treatment Interactions 119
Implications for Future Research 135Implications for Educational Practice 139Limitations and Further Statistical
Considerations 141
143APPENDICES
A. Raters Manual 144
B . Set Induction Materials . 154
157
D . Written Measures 162
E . Rating Forms 166
F. Intercorrelations Among Dependent
C. Written Model
Variables 169
Analyses For Homogeneity Assumptions . . 173
H. Analyses of Variance: ClassroomPerformance Measures 176
iv
I. Newman-Keuls Tables: ClassroomPage
Performance Measures 185
J. Newman-Keuls Tables: Written Measures 194
K. Intercorrelations of Ability Measures 197
BIBLIOGRAPHY 201
LIST OF TABLES
Table Page
1. Organization of Task and Ability Variablesin a Multi-Process Model of Learning . 19
2. Number of Sample Subjects by Sex andSubject Matter 34
3. Ability Measures 37
4. Treatment Procedures 41
5. Rater Reliability II 56
6. Independent and Dependent Measures 58
7. Intercorrelations Among DependentVariables for Total Sample 60
8. Means and Standard Deviations ofDependent Variables 62
9. Analysis of Variance of Analytic Questions 64
10. Summary of Analysis of Variance ofAnalytic Questions: Simple MainEffects 65
11. Analysis of Variance of Categories ofAnalytic Questions 68
12. Summary of Analysis of Variance ofCategories of Analytic Questions:Simple Main Effects 69
13. Analysis of Variance of High QualityAnalytic Questions 72
14. Summary of Analysis of Variance of HighQuality Analytic Questions: SimpleMain Effects
vi
73
Table Page
15. Analysis of Variance of Nonanalytic76Questions
16. Summary of Analysis of Variance ofNonanalytic Questions: Simple MainEffects 77
17. Analysis of Variance of Scores forTrue-False Test 81
18. Analysis of Variance of Scores forMatching Test 82
19. Means and Standard Deviations ofAbility Measures 83
20. Reliabilities and Standard Errors ofMeasurement for Ability Measures 85
21. Intercorrelations of Ability Measures:Total Sample 86
22. Correlations of Ability Measures withPerformance Measures: Film-MediatedModeling Group . . o 89
23. Correlations of Ability Measures withPerformance Measures: WrittenModeling Group 90
24. Correlations of Ability Measures withPerformance Measures: Control Group . 91
25. Simple Regression Analyses forAnalytic Questions 95
26. Simple Regression Analyses for Categoriesof Analytic Questions 98
27. Simple Regression Analyses for HighQuality Analytic Questions . . 101
28. Simple Regression Analyses:True-False Test 106
vii
Table
29. Simple Regression Analyses:Matching Test
viii
Page
110
LIST OF FIGURES
Figure Page
1. A Multi-Process Model of Learning 17
2. Mean Frequency of Analytic QuestionsAcross Teaching Sessions 66
3. Mean Frequency of Categories of AnalyticQuestions Across Teaching Sessions 70
4. Mean Frequency of High Quality AnalyticQuestions Across Teaching Sessions .. 74
5. Mean Frequency of Nonanalytic QuestionsAcross Teaching Sessions 78
6. Disordinal Interaction 93
7. Interaction of Hidden Figures--I Scoreswith Instructional Treatments forCategories of Analytic Questions 96
8. Interaction of Hidden Figures--I Scoreswith Instructional Treatments forHigh Quality Analytic Questions . . 99
9. Interaction of Hidden Figures--I Scoreswith Instructional Treatments forHigh Quality Analytic Questions
10. Interacticn of Film Memory Scores withInstructional Treatments for HighQuality Analytic Questions
11. Interaction of Film Memory Scores withInstructional Treatments for True-False Test Scores
12. Interaction of Maze Tracing--II Scoreswith Instructional Treatments forTrue-False Test Scores
13. Interaction of Maze Tracing--II Scoreswith Instructional Treatments forMatching Test Scores
ix
102
103
107
108
111
Figure Page
14. Organization of Ability Variables in a Multi-Process Model of Learning 122
7111,1,71,
SUMMARY
The purpose of this study was to examihe the ef-
fects of verbal and perceptual dimensions of individual
differences in relation to the efficacy of two different
kinds of modeling procedures in the acquisition of a teach-
ing skill. The two different methods of presentation em-
ployed were written and film-mediated models. Both model-
ing treatments were expected to produce greater behavior
change than a control treatment. It was anticipated that
for Ss receiving the film-mediated model, criterion scores
should show stronger relation to perceptual abilities, while
for Ss receiving the written model, criterion scores should
show stronger relation to verbal abilities. These hypothe-
ses imply that there are nonparallel regression slopes, and
consequently, that one treatment will not be superior
throughout the distribution of perceptual and verbal abil-
ity variables.
Specific predictions, though tentative, were
based on theoretical considerations which suggested that
the requirements of the Written and Film-Mediated Modeling
treatments were sufficiently different to produce different
xi
ability-performance relationship. These predictions were
derived from an analysis of learning tasks and processes
corresponding to a theoretical model proposed for the in-
vestigation of individual differences in learning.
Following the administration of aptitude tests
selec -ted- from the French Kit of Reference Tests for Cogni-
tive Factors and from a series of specially developed audio-
visual testing materials, 121 Stanford intern teachers were
randomly assigned to three treatment groups: a Film-
Mediated Modeling treatment, consisting of a filmed por-
trayal of Analytic Questioning, the teaching skill to be
learned; a Written Modeling treatment consisting of a ver-
batim text of the sound track from the film-mediated model;
and a Control treatment which received no model, but went
through all other steps common to the two modeling treat-
ments;
Treatment conditions were held constant in all
ways except for the mode of model presentation. In terms
of general procedures all Ss received the initial instruc-
tions and microteaching pretest followed by two cycles com-
prised of models, rehearsal and microteaching Treatments
were terminated with the completion of two written tests.
The criterion performances assessed included the frequency,
xii
variety) and quality of Analytic Questions used upon three
separate teaching sessions in addition to scores on the
written measures.
Instructional treatment main effects as well as ap-
titude X treatment interactions were investigated. Analy-
sis of variance was used to test treatment effects. These
analyses disclosed highly significant treatment effects in
which Ss in both Written and Film-Mediated Modeling treat-
ments generated significantly higher frequency, variety and
quality of Analytic Questioning than did Control group Ss.
Similarly, Ss in the Written and Film-Mediated Modeling
treatments performed significantly better on both written
measures than did Control group Ss. Moreover, from the
average data alone, training under Film-Mediated Modeling
conditions appears to have been consistently more effective
than training under Written Modeling conditions across all
measures of the dependent variable.
Aptitude x treatment interactions were evaluated
by comparing regression slopes for different treatments.
Analyses of interactions disclosed that scores on Hidden
Figures, Maze Tracing and Film Memory tests interacted sig-
nificantly with the instructional treatments. In addition,
the magnitude of these interactions tended to increase
across performance trials. Interpretations of these inter-
actions were derived within the framework of the theoreti-
cal model described.
These findings suggest that the rate and level
of learning of a specific teaching strategy varies as a
function of model presentation; and that the effectiveness
of instructional methods varies from S to S with such dif-
ferences being related to trainee aptitudes. Results such
as these, if replicated, may provide a basis for the indi-
vidualization of teacher training programs.
xiv
CHAPTER I
THE PROBLEM
Definition of the Problem
The fields of education and psychology
abound with questions related to the attainment of
learning proficiency through variations in materials
and tasks, instructional techniques, and abilities
of the individual learner. There are also the more
complicated questions of the degree to which learning
effectiveness is a function of the interaction of
these variables.
Investigations of learning under different
instructional procedures commonly assign Ss randomly to
multiple treatment conditions and compare average perfor-
mance on some criterion measure. The usual conclusion
has been that one of the treatments is more effective
1
2
than others in some general sense. However, the meaning
of main effects is altered in the presence of interactions,
and variables not represented have no opportunity to dem-
onstrate their interactive effects (Snow, Tiffin, and
Seibert, 1965).
Investigators have often looked for aptitude
variables to serve as general predictors of criterion
scores, but this tends to focus attention on aptitudes
that correlate with outcome under almost any treatment
rather than on those likely to be involved in aptitude X
treatment interactions. Predictors having low or negative
correlations with success under certain treatments have
been largely disregarded, even though these variables may
be of considerable importance when instruction is adapted
to the individual (Cronbach, 1965).
As Cronbach (1957) has argued:
Applied psychologists should deal with treatmentsand persons simultaneously. Treatments are charac-terized by many dimensions; so are persons. The twosets of dimensions together determine a payoff surface(p. 680) . . . Ultimately we should design treatmentsnot to fit the average person, but to fit groups ofstudents with particular aptitude patterns whichcorrespond to (interact with) modifiable aspects ofthe treatment (p. 681).
Gagn4 (1964) also has suggested that individual
differences in the learner are among the most important
independent variables as one proceeds to the more complex
forms of learning. To this extent he was supporting the
Cronbach argument for simultaneous consideration of aptitude
and instructional variables.
One complex form of learning in which the role
of individual differences might be studied is that of ob-
servational learning. The wealth of literature surrounding
observational learning (Bandura and Walters, 1963; McDonald
and Allen, 1967) indicates the educational significance of
studying learning efficiency in association with modeling
procedures. Research in this area has consistently shown
that complex social responses and teaching skills may be ac-
quired or the characteristics of response hierarchies may be
considerably modified as a function of observing the behav-
ior of others and the consequences of their responses, with-
out the observer performing an overt response himself, or
receiving any direct reinforcement during the acquisition.
While Bandura and Walters (1963) have suggested
that observer characteristics influence the extent to which
observational learning occurs, observer characteristics
have been largely disregarded in previous research. Thus
a study of the relationship between individual differences
and the efficacy of different modeling procedures, while
4
exploratory, contains elements of potential importance
for evaluating the factors influencing learning efficiency
when modeling procedures are used in teacher training
programs.
Related Theory and Research
Observational. Learning
A common approach to the transmission or modifi-
cation of teaching skills has been to provide teacher
trainees with some type of written or oral instruction
followed by periodic feedback on their attempts to produce
these behaviors in the classroom. An alternative to this
strategy of teacher training is suggested by recent find-
ings on the role of observational learning in personality
development and behavior modification. Bandura and Walters
(1963), in a review of the relevant literature point out
that complex behavior may be acquired almost entirely
through imitation. The provision of live or symbolic
models serves to accelerate the learning process, and,
in cases where time or error is costly, this technique
becomes an efficient means of transmitting behavior pat-
terns (Bandura and Walters, 1963; Bandura, Ross and Ross,
1963). Other research has shown that the level of the
desired behavior exhibited by the observer can be as great
as that exhibited by the model (McBrearty, Marston, and
Kanfer, 1961). The implication of these findings for
teacher training is that the provision of live or symbolic
models displaying desired teacher behaviors may provide
an effective alternative to purely descriptive techniques
of training, (McDonald and Allen, 1967).
The concept of imitation in psychological theory
has a long history dating back to Lloyd Morgan (1896),
Tarde (1903) and the doctrine of instinct. In more recent
formulations it has been widely assumed that the Occurrence
of imitation, or observational learning, is contingent
upon the administration of reinforcement either to the
model or to the observer.
According to the theory proposed by Miller and
Dollard (1941), the conditions necessary for learning in-
clude a motivated subject who is positively reinforced
for matching the rewarded responses of a model.
Mowrer's (1960) proprioceptive feedback theory
similarly highlights the role of reinforcement, focusing
on the classical conditioning of positive and negative
emotions to matching response correlated stimuli. According
6
to Mowrer, imitative learning can occur either through
pairing of the model's response with reinforcers dispensed
to the observer, or by means of vicarious conditioning.
However, these reinforcement theories do not
account for the learning of matching responses when the
subject does not perform the model's responses during
the process of acquisition and for which reinforcers are
given neither to the model nor to the observer. The ac-
quisition of imitative responses under the latter condi-
tions can best be explained in terms of contiguity theory
of observational learning (Bandura, 1965; Sheffield, 1961).
According to this theory, it is assumed that when an ob-
server witnesses a model exhibiting a sequence of responses,
the subject acquires through the contiguous association
of sensory events, symbolic or representational responses
possessing cue properties capable of eliciting at a later
time, overt responses matching those that have been pre-
viously observed. Accordingly, Bandura (1963, 1965) has
argued that contiguity accounts for the acquisition of
matching responses, whereas reinforcement influences the
performance of imitatively learned responses.
It cannot be assumed that observational learning
is contingent solely upon exposure of an observer to a
7
complex sequence of stimulation. Bandura (1965) states
that imitation is an active process in which modeling
stimuli combine with other variables in shaping response
patterns. Factors other than contiguity undoubtedly in-
fluencing imitative response acquisition would include:
attention directing variables such as motivation, prior
training in observation, and anticipation of reinforcement;
rate, amount and complexity of stimuli presented to the
observer; practice; observer charanteristics and other
variables.
McDonald and Allen (1967) have applied the re-
search on observational learning to the learning of teach-
ing behavior, and have obtained data pertaining to the
following independent variables: (1) self-feedback and
reinforcement; (2) variations in feedback and practice
conditions; (3) film-mediated and written modes of model
presentation. These studies have indicated that reinforce-
ment and discrimination training administered by an ex-
perimenter were highly effective methods of producing
behavior changes in teachers. Because of considerable
information processing capabilities, human learners nay
require little in the way of feedback under many circum-
stances. However, if the responses to be learned are
8
sufficiently complex, some type of feedback may be required.
Moreover, exposure of an observer to a complex sequence
of teaching behavior does not guarantee that Ss will attend
to the most relevant stimuli, or accurately perceive the
cues to which their attention is directed. Discrimination
training, in which attempts have been made to focus Ss
attention on relevant stimuli, has been shown to facilitate
observational learning.
In the preceding experiments, it was observed
that different kinds of training procedures appear to be
differentially effective with trainees. While a great deal
of individual variation has been observed within the written
and film-mediated modeling conditions, differences between
the two modes of model presentation have not been consis-
tently significant. A promising line of research suggested
by these results would appear to be a study of the inter-
actions between the kinds of treatment used to produce
behavior change and (1) the type of teaching skill to be
learned, and (2) individual differences in the learner.
A study has recently been undertaken to investigate inter-
actions between various modeling procedures and types of
dependent variables (McDonald and Allen, 1967). The pur-
pose of the present study was to investigate the interaction
9
of modeling procedures with individual differences in the
learner.
Learning and IndividualDifferences
The relationship of tested abilities to learning
rate has been of experimental interest for quite some
time. Early investigators have examined the generality
of learning ability to determine if a single learning
ability accounts for performance in all learning situations.
Over a wide variety of learning tasks, it has generally
been found thr.t measures of learning in such tasks exhibit
low correlations both with other tasks and with intelli-
gence measures. The usual conclusion has been that there
is no general learning ability.
Husband (1939, 1941) was among the first to
conclude that learning ability is not unitary but that
there are a number of relatively independent learning
abilities depending upon the type of task. The most sub-
stantial evidence on this point, however, comes from a
series of studies by Woodrow (1938, 1939, 1940, 1946) in
which a group of students were given a number of practice
trials on a variety of learning tasks. A battery of special
10
aptitude and intelligence tests was administered both before
and after practice. The results obtained showed that there
was no single general learning ability accounting for im-
provement on all tasks. Moreover, there was no significant
relationship between scores representing gain with practice
and intelligence. Additional research in this area (Simrall,
1947) has indicated that even when the types of material
used in the learning tasks were highly similar to types
of material used in the intelligence tests, gains during
practice were still not significantly related to test
scores. While methodological weakness and inappropriate
statistical analysis (Rapier, 1962, Manning and Dubois,
1962) limit the usefulness of these early studies as a
basis for conclusions, later interpretations (Humphreys,
1960; Porter, 1959) have generally supported the position
that neither the amount gained nor the rate at which it
is acquired show any consistent relationship to intelligence
measures.
Related to the Woodrow trend is the work of
Fleishman and his associates (summarized in Fleishman,
1965) on the relationship of ability factors to performance
in the course of perceptual-motor skill learning. In-
vestigators of learning have commonly used a terminal
11
performance score as a criterion for learning. Such a
score is believed to be an accurate representation of the
learning that has taken place. However, Fleishman's find-
ings have suggested that the learning of complex skills
is a multidimensional process in which the contributions
of individual differences associated with performance on
a task change from early to later stages of learning.
These findings serve to underscore the importance of ex-
perimental work on behavioral change which investigates
the effects of task variables, particularly as they change
over conditions of practice or instruction, (Glaser, 1967).
The line of work generated by Woodrow is reflected
today in the work of other differential psychologists who
have related aptitudes to performance on various laboratory
tasks (Allison, 1960; Stake, 1961;Duncanson, 1964). Gen-
erally, in these studies, a battery of reference tests is
administered, the subjects complete various learning tasks,
and regression or factor analytic techniques are used to
relate learning to the reference tests. However, this
line of investigation has been primarily concerned with
task variables rather than instructional variables. While
these studies have investigated the relationship of apti-
tudes to performance under conditions of practice, they
12
have not generally compared ability-performance relation-
ships under different methods of instruction. Thus, these
results typically do not indicate any specific way in which
Ss could be treated or situations modified in order to
maximize learning.
Direct pursuit of interactions between treatment
and learner variables is infrequently represented in the
literature. While interactions with instructional variables
have most commonly been reported.as incidental findings
in experiments designed for other purposes, there are a
few studies that have deliberately investigated relation-
ships among learner characteristics and learning outcomes
under different instructional conditions.
An investigation of individual differences
and instructional film effects (Snow, Tiffin and Seibert
1965), using filmed and live lecture demonstrations found
that a number of variables including ascendancy, responsi-
bility, numerical aptitude, verbal aptitude, past experience
with entertainment films, and past use of college library
instructional films interacted significantly with instruc-
tional treatments.
Edgerton (1958), using Navy training courses,
taught the same material by two different methods in an
13
effort to modify the conventional rote instruction into
a meaningful method. Correlations obtained indicated that
the Thurstme memory and fluency factors interacted with
the instructional conditions.
Grimes and Allinsmith (1961) compared primary
reading achievement under two methods; a structured phonics
program, and a less structured whole-word approach, with
anxiety and compulsivity used as differential variables.
It was anticipated that highly anxious or highly compulsive
children would show more achievement under structured
methods than similar children taught by the unstructured
method. This expectation was supported. Moreover, anxiety
and compulsivity, which were not correlated in their sample,
interacted with each other as well as with instructional
method.
Finally, Cronbach (1965), in a partial survey
of the literature dealing with attitudes concerning con-
fidence, willingness to risk failure, and motivation for
self directed achievement, examined numerous indications
from the work of Atkinson (1964), Wallach and Kogan (1964)
and others that certain personality variables interact
with instructional variables.
It seems clear that interactions may be dis-
covered in a random exploration in which results under
14
multiple instructional treatments are projected onto a
mass of differential information. Cronbach (1965), however,
argues for what he considers a more promising experimental
strategy in which alternative treatments are designed to
interact with specific differential variables. The design
of the present study corresponds more closely to the latter
strategy in that treatments were initially selected in
terms of specific aptitude variables and then refined to
maximize the effects of these variables.
Implications of Research onObservational Learning andIndividual Differences
What are the implications of the previous dis-
cussion for study of the interaction of individual differ-
ences with modeling procedures in the acquisition of a
teaching skill?
First of all, instructional methods differ. The
literature reviewed here suggests that a person learns
more easily from one method than another, that this best
method differs from S to S and that such differences be-
tween treatments are related to learner characteristics.
Different methods of presentation may be employed
in the use of modeling procedures. That is to say that
15
different methods of presentation may be used to reach
the same terminal objective. Among these are written and
film-mediated models. For the purposes of this study,
by "exposure to a film-mediated model" is meant that the
learner has observed the actual filmed performance of
another person who displayed the behaviors to be acquired.
By "exposure to a written model" is meant that the learner
has read a written transcription of the sound track from
the film-mediated model. While both of these methods
have been effective as training procedures, differences
between the two treatments have not consistently been
found (McDonald and Allen, 1967; Bandura and Mischel,
1965). However, in view of the task differences generated
by the two different modeling treatments, it seems reason-
able to expect that different abilities may also be in-
volved.
A study by Wolfgang Boehm (cited by Cronbach,
1957) tends to lend support to this expectation. In this
study, a sound film was shown to the experimental groups.
Matched control groups read a verbatim text of the sound-
track. Differences in comprehension between the two groups
were insignificant. However, a general mental test cor-
related only .30 with text learning, but predicted film
16
learning with an average correlation of .77. This dif-
ference was consistent across all ages and would appear
to support the hypothesis that different abilities might
predict success on these two different methods of presenta-
tion.
In answer to the question of why purely audio-
visual and purely written verbal treatments should be used
when a combination of the two might be likely to produce a
stronger training effect for all, or for any one or more
sub-groups, it is suggested that such a separation is a
useful research strategy which may offer a clearer under-
standing of the separate functions of the treatments and
abilities involved. A mixed treatment might mask the
functioning of aptitudes which may be interacting with
treatment effects. Moreover, the two treatments utilized
here have already been defined as meaningfully distinct by
previous research. These treatments have both theoretically
and practically distinguishable features. As Cronbach (1957)
has argued, unless one treatment is clearly best for every-
one, a situation which has not occurred for the treatments
under consideration here, treatments should be differen-
tiated in such a way as to maximize their interaction with
aptitude variables. In this way, if interactions do exist,
learning is maximized.
17
Definition of Relevant Abilities
While individual differences in performance on
learning tasks can be measured as a function of numerous
operationally defined variables in those tasks, and factor
analytic studies can relate performance on a number of
tasks to a variety of reference tests, ultimately, research
on individual differences must be guided by theories of
human learning and performance. As Melton (1967) has argued:
What is necessary is that we frame our hypothesesabout individual differences variables in terms of
process constructs of contemporary theories of learn-
ing and performance (p. 239) .
Accordingly, Melton has proposed an expanded
version of a theoretical model initially described and
tested by McGuire (1961) as a framework for investigating
individual differences in learning. The model is shown
in Figure 1.
Figure 1. A Multi-Process Model of Learning
18
This model proposes a stimulus differentiation
component, a response integration component, and a "hookup"
between the encoded stimulus and the integrated response
or a segment of it. The stimulus differentiation com-
ponent (S1 ri) is the subjects' coding response to
the potential physical stimulus (SO. It is the effec-
tive, or functional stimulus (ri(si)) in S-R associations,
and reflects such factors as learning and set as well
as the physical stimulus. The response integration com-
ponent (RaRbRc) is the output response, which may be
either a previously learned unit (Ra) or a new combination
of such units, (RaRbRc). The "hookup" is the connection
(ri(s1)---4wRaRbRe) between the functional stimulus and
the required response. The component (rm(dm)') represents
an alternative mediational route for the connection of
the internal representation of the physical stimulus and
the required response.
Differences and similarities in task and ability
variables involved in the two modeling procedures previously
described may be represented in terms of this paradigm.
This representation is summarized in Table 1.
First, Ss in the Film-Mediated Modeling condition
view an actual portrayal of the teaching skill and of the
AIN
19
TABLE I
ORGANIZATION OF TASK AND ABILITY VARIABLES
IN A MULTI-PROCESS MODEL OF LEARNING
S1 _IN, ri( si) R aRb Rc
StimulusDifferentiation
AssociationMediation
ResponseIntegration
WrittenModel
Read script of soundtrack
Generate associativecontext
Set own reading pace Transform verbal tobehavioral repre-sentation
Reread, pause asneeded
Retain representation
Generate new associates
Integrate in behavioralcontext
Code as verbalrepresentation
View film, hearsound track
Generate associativecontext
Retain representations
Film-MediatedModel
Keep pace with film Transform behavioralto verbal repre-sentation
Generate new associates
Review in memory Integrate in behavioralcontext
Flexibility ofclosure
Verbal BehavioralTransformation
Verbal memory
Ability Perceptual speed Verbal Association Audio-Visual memory
VariablesSpatial scanning Verbal fluency
Short-term memory
Verbal comprehension
20
pace of the lesson. This involves processing information
from multiple channels simultaneously. Film-mediated
models are extremely rich in perceptual detail in that
there are many different cues to which to attend, including
many extraneous to the relevant features of the model.
These task characteristics would appear to require the
ability to keep definite task-relevant dimensions in mind
so as to make identification in spite of perceptual dis-
traction. This ability may be similar to that commonly
called flexibility of closure (Thurstone, 1944).
Moreover, Ss in the Film-Mediated Modeling condi-
tion are expected to perceive and encode events as be-
havioral representation at the speed of presentation of
the film. Consequently, they are able to review relevant
material in memory only. The requirements of a predeter-
mined pace would seem likely to involve speed in perceptual
evaluation and in exploring a complicated spatial field,
as well as short term memory facility.
In the Written Modeling condition, Ss read a
written transcription of the soundtrack from the film-
mediated model. A written transcript of the model's be-
havior, however, presents only verbal components of these
stimuli. Consequently, Ss in the Written Modeling condition
23.
process information from a single channel in which they
are able to establish their own pace and return to review
relevant material as read. While Ss in the Written Modeling
condition must also distinguish the significant features
of the model, the total set of stimuli to select from is
limited to verbal components. These task variables would
be expected to require the ability to comprehend written
verbal material.
Corresponding to the associational-mediational
component of the model, it is hypothesized that Ss in both
Film-Mediated and Written Modeling conditions must generate
their own analogy to the model's performance. While Ss
in the Film-Mediated Modeling condition are expected to
abstract relevant verbal representation from behavioral
representation, Ss in the Written Modeling condition are
expected to generate relevant behavioral representation
from solely verbal components. These task variables would
appear to require facility in both verbal association
and in verbal-behavioral transformation. Although these
abilities would appear to be involved in both treatments,
they are presumed to be more decisively involved in the
Written Modeling condition. Because of the abstraction
of the script, or compression of stimuli into solely verbal
22
components, a greater degree of verbal processing would
appear necessary for those in the Written Modeling
condition to generate associative context and behavioral
representation. Ss in the Film-Mediated Modeling condition
have only to recall much of what Ss in the Written Modeling
condition must produce.
Finally, corresponding to the response integration
component of the model, it is expected that Ss in both
conditions need to retain representations of their respec-
tive models, generate new examples of the skill being
modeled in the teaching situation, and integrate these
examples into a behavioral context. These task variables
would be expected to require both verbal fluency, and the
ability to remember major ideas as well as details from
audio-visual presentations in the case of Ss in the Film-
Mediated Modeling condition, and verbal presentations in
the case of Ss in the Written Modeling condition.
Ability Measures
The selection of ability measures for use in this
study was made from the Kit of Reference Tests for Cognitive
Factors (French, Ekstrom and Price, 1963) and from a series
of film and audio tests developed by Seibert and Snow (1965)
23
and Seibert, Reid and Snow (1967). The abilities assessed
are believed to be those which distinguish most clearly
between the two modeling conditions and are consistent
with the analysis of the Film-Mediated and Written-Modeling
conditions previously discussed.
The Kit of Reference Tests for Cognitive Factors
was developed from a project organized to select tests to
represent each of the better established factors in the
cognitive behavioral domain. It consists of a group of
tests representing each of the more frequently obtained
factors in the cognitive ability area, rather than a stan-
dardized battery of tests. The purpose of the kit is to
provide research workers with a set of tests for defining
each of these factors and for facilitating interpretation
and comparison of one factor study with another.
The manual does not provide the reliability,
forming, validity or other information usually included
in a test manual. Such information has not been included
because these tests are suggested for the single purpose
of factorial research. Consequently, there is no compre-
hensive information on how these tests intercorrelate.
Where there is information on some of the tests, it has
been gathered on samples dissimilar to the one used in
24
this study. However, all of the tests have two separately
timed parallel parts to permit the estimation of reliability,
and a correlation matrix has been computed to provide
additional necessary information for subsequent discussion.
The film and audio tests were developed in an
effort to examine alternatives to conventional printed
paper tests in psychometric research. It is argued by
Seibert and Snow (1965) that since human beings behave
in contexts which are not static, not primarily verbal,
and not characterized by the presence of print, it is
conceivable that some of this behavior involves abilities
or predispositions which are not amenable to adequate
sampling with printed paper tests.
The purpose of their investigation was to begin
the exploration of individual differences in performance
on tasks characterized by moving, sequential or behavioral
content. Since those characteristics were involved in
the present study, it was thought that these tests would
provide useful information. Tests chosen from the Kit of
Reference Factors for use in this study have also appeared
prominently in the factor interpretations of the earlier
film test study.
25
A brief description of the ability tests and
the factors they represent follows:
Hidden Figures (Cf-1, Flexibility of Closure).
A test of the ability to keep one or more definite con-
figurations in mind so as to make identifications in spite
of perceptual distraction. The visual and auditory modali-
ties of the closure factor have been shown to be highly
correlated. (White, 1954)
Identical Pictures (P-3, Perceptual Speed).
A test of speed in finding figures, making comparisons
and carrying out other tasks involving rapid checking
and perceptual evaluation.
Maze Tracing (Ss-1, Spatial Scanning). Related
to perceptual speed is spatial scanning ability, which
is described as speed in visually exploring a wide or
complicated spatial field. Tests of perceptual speed
and spatial scanning tended to appear together on a film
test factor (Seibert and Snow, 1965) hypothesized to in-
volve rapid serial processing of spatial images, and sug-
gesting the dynamic information processing nature of this
ability.
26
Advanced Vocabulary (V-4, Verbal Comprehension).
A test of the ability to understand the English language.
Individual differences are most clearly seen in the size
of comprehension vocabularies, but also exist with respect
to tests demanding knowledge and understanding of grammati-
cal patterns, sentences, phrases and other aspects of the
English language.
Word Arrangement (Fe-3, Expressional Fluency).
A test of ability to think rapidly of appropriate wording
for ideas. The emphasis in this test is on facility in
producing connected discourse that will fit restrictions
imposed in terms of given words.
Film Memory. A test of the ability to iden-
tify verbal descriptions from remembered behavioral con-
tent. It consists of a motion picture presentation,
followed by printed questions concerning the action por-
trayed.
Memory for Ideas. A test of ability to ab-
stract major ideas from remembered audio taped verbal
presentation.
27
Sentence Reproduction. A test of the ability
to reproduce verbal material in detail from remembered
audio taped verbal presentation.
GRE Verbal Aptitude scores were also used in
this study. This score represents verbal reasoning and
reading comprehension.
The Dependent Variable
Educators hold that one of the major objectives
of instruction in the schools is the development of skills
or strategies which will enable students to become life-
long autonomous learners. While the content of any field
may undergo rapid change, the ability to analyze crit-
ically written materials within that content area is
an objective of any field of study (Bloom, 1956). One
way of developing such ability in students is to train
teachers to ask the kinds of questions that require stu-
dents to engage in the analysis of their reading material.
With this objective in mind, the dependent variable de-
veloped for use in this study is termed Analytic Ques-
tioning.
The basic ideas underlying the development of Ana-
lytic Questioning_ were derived from Taxonomy of Educational
28
Objectives (Bloom) 1956) which presents a plan for classi-
fying educational objectives. This taxonomy was developed
from a project organized to achieve a more widely accepted
set of behavioral objectives that could be more profitably
used for communication and for guiding research in connec-
tion with curricula, teaching and examining in education.
Bloom and his associates claim that any educa-
tional objective can be classified within their taxonomy,
and imply that any question can also be classified. One
of the ways in which each category was defined was by using
examples of questions that required students to engage in
specified kinds of behavior (Sanders, 1966). Accordingly,
if the objectives of the taxonomy are considered to be
worthwhile educational objectives, training teachers to
formulate questions requiring students to engage in the
behavior specified as an objective would consequently be
considered a worthwhile training objective for prospective
teachers. Thus, analysis of elements, a subset of the
general objective of Analysis, was selected as an appropri-
ate source for training teachers in questioning behavior.
The Taxonomy of Educational Objectives defines
Analysis in the following manner:
29
The breakdown of a communication into its constituentelements or parts such that the relative hierarchy ofideas is made clear and/or the relations between theideas expressed are made explicit. Such analyses areintended to clarify the communication, to indicate howthe communication is organized, and the way in whichit manages to convey its effects, as well as its basisand arrangement (p. 205)
Although Analysis may be conducted simply as an
exercise in identifying structure and organization in a
communication, it is probably more educationally useful
as an aid to fuller comprehension or as a prerequisite
to higher level skills in the taxonomy, such as Synthesis
or Evaluation.
The Analysis of Elements is specifically con-
cerned with the breakdown of a communication into its
constituent parts, i.e. to identify or classify the elements
of a communication. Any communication may be conceived
of as composed of a number of elements, some of which are
e xplicitly stated in the communication and can be recognized
and classified relatively easily. However, there are other
e lements in a communication which are not as clearly labeled.
These elements can be inferred only from an analysis of
a series of statements within the communication. Until
the reader can detect them, he may have difficulty in fully
comprehending or evaluating the communication.
30
A detailed discussion of the dependent variable
appears in the Raters Manual (Appendix A). Thus comments
here will be brief. The categories of Analytic Questioning
are the following:
1. Identification of Hypotheses
2. Semantic Definition
3. Identification of Unstated Assumptions
4. Distinction of Factual from Non-Factual
Statements
5. Identification of Conclusions
The labels in each case generally reflect the teacher's
goal when using a given category of questions.
Statement of Hypotheses
Based upon the previously reviewed research
and theory, the basic hypotheses derived and tested were:
1. Both Film-Mediated and Written Modeling
conditions will produce significantly greater
changes in the response strength of desired
behaviors than will the Control condition
2. For subjects receiving the Film-Mediated
Modeling condition, criterion scores should
show stronger relation to perceptual abilities
31
thEn for subjects receiving the Written
Modeling condition.
3. For subjects receiving the Written Modeling
condition, criterion scores should show
stronger relation to verbal abilities than
for subjects receiving the Film-Mediated
Modeling condition.
These hypotheses imply that there are nonparallel
regression slopes, and consequently, that one treatment
will not be superior throughout the distributions of the
perceptual and verbal ability variables. For the purpose
of this study the term "perceptual abilities" has been
used to include both tests of figural cognition from the
French Kit and the audio visual memory tests. While the
audio-visual tests are experimental in nature, and it is
not yet clear precisely how they should be classified,
they have been tentatively grouped with the tests of fig-
ural cognition, mainly because of their mode of presenta-
tion. The term "verbal abilities" has been used to include
written tests of verbal cognition.
Although perceptual and verbal abilities are
referred to here in a seemingly unitary sense, it should
be emphasized that tests representing these abilities
32
are presumed to be independent factor tests. While it
may be anticipated that both perceptual and verbal abili-
ties would display a tendency to correlate more highly
with one another than with other more dissimilar measures,
identical ability-performance relationships have not been
anticipated for all tests of a given class. The relation-
ship of any particular test labeled "perceptual" or "verbal"
to any particular instructional outcome may be expected
to vary as a function of its relevance to both task and
process variables.
CHAPTER II
EXPERIMENTAL DESIGN
Subjects
Intern subjects were drawn from the Stanford
Teacher Education Program. They were initially categorized
by the subject matter they were teaching, and subsequently,
randomly assigned to one of three treatment conditions.
All of the subject matter areas of the Stanford Teacher
Education Program were represented with the exceptions of
foreign language and physical education majors. In both
of these cases it was felt that the prescribed methodology
was inappropriate for the systematic use of the dependent
variable, Analytic Questioning. Moreover, use of the audio
tapes of foreign language majors would have created undue
transcription and measurement problems.
Relevant characteristics of the sample studied
appear in Table 2 which provides the distribution of sub-
jects by sex and subject matter.
The Ss ranged in age from 21 to 40 years, al-
though the variability was relatively small with most of
33
TABLE 2
NUMBER OF SAMPLE SUBJECTS BY SEX AND SUBJECT MATTER
N = 121
Subject Matter Mai or
TreatmentGroup N
Sex
4m..-1
1-4
too
W
wr-i 0as o.0-1 WC) rl0 0ca ca
w0oW
4-10co
1 ti
cu 04-H+3 -I-)cd cdX M
M F
Film-MediatedModeling
WrittenModeling
Control
41
40
40
8
9
12
33
31
28
13
11
11
17
18
16
5
4
5
3
3
3
0..-1
+3k
m0
cg X
2 2
2 3
2 1
35
the ages falling near the mean of 23.0. The mean under-
graduate Grade Point Average of this group was 3.03.
Interns taught each of their lessons to a micro-
class of four students. These students were randonly as-
signed to teams of four which were subsequently randomly
assigned to Ss within treatment groups. The microstudents
were drawn from the Palo Alto school system and were paid
for their services. A total of 120 students were hired.
All students were either tenth or eleventh graders with
a grade point average of 3.0 or better. The majority of
these students were from middle class backgrounds.
Ability Measurement
Ss were tested prior to their participation in the
experimental procedures. Tests were administered in a
single two and one-half hour session on two separate oc-
casions. Testing time included instructions, practice
exercises and a fifteen minute rest interval.
In the first testing session, it was impossible
for all Ss to be tested together in the same room because
of the size of the intern group and the nature of the
36
audio-visual tests. Consequently, the one large group was
divided into two smaller groups which were tested simul-
taneously, in separate rooms. Moreover, the amount of
equipment involved in the administration of the audio-
visual tests, and the limitations imposed by physical
facilities did not permit the duplication of audio-vi;i,w1.
facilities in both rooms. Therefore, each group was ad-
ministered the tests in two separate sets--a set of audio-
visual tests and a set of paper and pencil tests from the
Kit of Reference Tests for Cognitive Factors. Upon com-
pleting the first set of tests, subjects changed rooms
and completed the second set. Within each of these two
sets the order of test administration was randomly assigned.
A second testing session was provided for those
Ss unable to attend the first session. This group was
small enough to permit all Ss to be tested together. Thus,
in this group it was possible to administer the two sets
of tests in a random mixed order. A summary of the testing
procedures appears in Table 3.
Tests were hand scored by three trained scorers
working under the direction of the experimenter. Objec-
tive test scoring keys were used on all tests with the
exceptions of Word Arrangement, Memory for Ideas and
TABLE 3
ABILITY MEASURES
*N.
NumberTest Time of
Parts
Presentation Order
Session One Session Two1 2 1
N=45 N=42 N=34
Kit of Reference Tests forCognitive Factors
Hidden Figures (Cf-1) 20 min 2 1 4 7
Identical Pictures (P-3) 3 min. 2 2 5 8
Maze Tracing (Ss-1) 6 min. 2 3 6 3
Word Arrangement (Fe-3) 8 min. 2 4 7 2
Advanced Vocabulary (V-4) 8 min. 2 5 8 4
45 min.
Audio-visual Tests
Film Memory 7 min. 1 6 1 1
Memory for Ideas 16 min. 2 7 2 5
Sentence Reproduction 34 min. 2 8 3 6
57 min.
38
Sentence Reproduction. Scoring directions for these tests
were sufficiently unambiguous as to cause little scoring
uncertainty. When such uncertainty did arise, it was
resolved by the experimenter. Scoring was checked for
accuracy. Both part scores and total scores were recorded
for each test. GRE-Verbal Aptitude scores were available
from Stanford Teacher Education Program records.
Treatment Procedures
A total of 121 Ss were used in this study. All
Ss had been administered the selected ability measures
prior to the presentation of the learning task. Each S
was randomly assigned to one of three treatment conditions:
a Film-Mediated Modeling condition in which subjects were
exposed to filmed portrayal of the behavior to be learned,
Analytic Questioning; a Written Modeling condition in
which Ss read a verbatim text of the sound track from the
film-mediated model; and a Control group which received no
model, but went through all other steps common to the
previously mentioned two groups.
The treatment procedure may be broken down into
nine steps. In seven of these steps, all groups received
39
identical treatment. In the remaining two, each group
was exposed to the appropriate type of modeling procedure.
Thus, treatments were held constant in all ways except
for the mode of model presentation.
The treatment procedure was broken down into
the following steps:
1. Set Induction--Ss were presented with brief writ-
ten instructions describing the nature of the
learning task (Appendix B).
2. Teaching Session One, (Ti)--A preevaluation of
each S was made prior to the presentation of the
model as each intern presented a ten minute les-
son to a microclass of four students. Criterion
scores from this lesson provided a baseline fre-
quency of the desired behavior.
3. Model Lesson One--Depending on their assignment
to an experimental condition, Ss were exposed to
a written or film-mediated version of a model
exhibiting Analytic Questioning behavior. The
Control group was given a set of written materials
extr= aneous to the learning task in order to fill
the equivalent time block,.(Appendix C).
40
4. Rehearsal One--Ss were given five minutes to in-
corporate the modeled behavior into their planned
lesson.
5. Teaching Session Two, (T2)--Ss taught a second
ten-minute lesson.
6. Model Lesson Two--Ss were exposed to the same
model presentation as used in step 3.
7. Rehearsal Two--Ss were given five minutes to in-
corporate the modeled behavior into their planned
lesson.
8. Teaching Session Three, (T3)--Ss taught a final
ten-minute lesson.
9. Testing Session--Ss were given two written tests
covering material presente,1 in the treatment pro-
cedures (Appendix D).
A summary of the procedures appears in Table 4.
All Ss thus received written instructions followed by two
cycles comprised of models, rehearsal, and teaching. Treat-
ments were terminated with the completion of the written
tests. Ss taught a different group of students in each
lesson. This permitted them to revise the same basic sub-
ject matter while attending to the improvement of Analytic
Questioning behavior.
TABLE 4
TREATMENT PROCEDURES
Steps GroupTime
Film-MediatedModeling
Written Modeling Control
1. Set Induction X X x 5 min.
2. Teach 1 X X X 10 min.
3. Model 1 film mediated written extraneousmaterial
10 min.
4. Rehearsal 1 X X X 5 min.
5. Teach 2 X X X 10 min.
6. Model 2 film mediated written extraneousmaterial
10 min.
7. Rehearsal 2 X X X 5 min.
8. Teach 3 X X x 10 min.
9. Test Administration X X x 5 min.
Explanation of Symbols: (X) indicates that all Ss received thisstep of the treatment in an identical manner. Written descriptionsare provided for the two steps in which treatments varied amongthe three groups.
42
Procedures for Model Preparation
An experienced teacher was selected to portray
Analytic Questioning in the film-mediated model. The
model was thoroughly familiarized with both the skill to
be modeled and the desired quality level through written
materials, discussion, and illustration of representative
examples from each category of Analytic Questioning. A
ten-minute lesson was planned by the model and experi-
menter together to provide at least two representative
examples of each category of Analytic Questioning. Micro-
students participating in the videotaping of the model
were provided training along with the teacher on the cate-
gories of questions to be asked, and appropriate responses
to each of these.
Several trials were provided for taping the
model lesson. During each trial, the experimenter recorded
the number and quality level of the Analytic Questions
used, and noted suggestions for improvement. The model
and experimenter would subsequently review the lesson
previously taught and make revisions where necessary.
From this pool the best of the tapes was selected.
Upon completion of the film-mediated model, a
brief commentary was superimposed on the soundtrac to aid
43
the discrimination of important behaviors in the lesson.
A verbatim text of the soundtrack was then typed from the
film-mediated model to provide the written model (Appen-
dix C) . The commentary superimposed on the soundtrack
of the film-mediated model appeared on the written model
in exactly the same form and place in the lesson.
Lesson Material
Because the extent to which Analytic QuestioniRE
can be systematically used is partially a function of the
opportunities provided for such questioning by the lesson
material at hand, lesson material was preselected for Ss
in this experiment. With the help of both faculty members
and intern supervisors, material was selected to meet the
following criteria:
1. The lesson material had to provide a rich source
for the systematic use of Analytic Questioning.
2. The lesson material had to be adaptable to the
ten-minute teaching interval established.
3. The content of the lesson material could not pre-
suppose a substantial amount of prior information
on the part of the microstudents.
44
A total set of thirteen written communications
was finally selected to serve as lesson material. The
number of lessons selected in each subject matter area
represented a ratio of approximately one lesson for every
eight interns in that subject matter area. This was done
to avoid overexposure of microstudents to any given lesson,
and also to demonstrate the utility of Analytic Question-
ing, as a questioning strategy, over a variety of materials.
The topics represented in this selection included student
activism, the hippie culture, the use of drugs, the defi-
nition of patriotism, the cultural value of pop music,
and others. Sources for this material included newspaper
and magazine articles in addition to material especially
written for the purposes of this study.
Lessons were randomly assigned to Ss within sub-
ject matter areas and treatments, and were distributed to
them one week prior to their participation in the experi-
ment. Interns were instructed to come prepared to teach
a ten-minute lesson discussing this written material with
the microstudents. In addition, they were told that the
microstudents would also have copies of this material which
they would be prepared to discuss at that time.
Procedures for Training Microstudents
45
Training was provided for microstudents for the
following reasons: (1) the complexity of the dependent
variable, (2) the ceiling effect on its frequency of oc-
currence imposed by time and availability of appropriate
student responses.
Treatments were designed to produce increases
in Analytic Questioning behavior. However, both the ques-
tioning strategy and the lesson material selected were
complex. While the ability level of the microstudents
was specified with this in mind, the questioning strategy
is tied to the analysis of the written material at hand.
Unless student responses were highly available, much of
the tenminute lesson period could have been spent in
searching the written material for appropriate answers,
limiting the frequency with which such questions could
be asked by the teacher. This would serve to mask the
functioning of treatment effects. Accordingly, optimal
conditions for the demonstration of treatment effects
would necessarily include high response availability on
the part of the students.
Training of the microstudents was accomplished
in the following manner. One week prior to the experiment,
46
all microstudents attended an orientation meeting in which
copies of each set of lesson materials to be used for
analysis were distributed along with a description of the
categories of questions which they could expect to be asked
with respect to that material. Examples of appropriate
questions and answers for each lesson corresponding to
each of the categories of Analytic Questioning were pro-
vided. Students were instructed to be well prepared on
this material as their participation in microteaching
would be contingent upon their successful completion of
a written test on the material to be administered prior
to each day of microteaching in which they were to par-
ticipate.
A random sample of four lessons was selected as
a source for this test prior to each day of microteaching.
Students were instructed to list at least one example of
material corresponding to each of the categories of Ana-
lytic Questioning for each lesson. These tests were checked
by the experimenter together with assistants. No micro-
student failed to meet these minimal criteria.
47
Criterion Measures
Consideration was given to several aspects of
Analytic Questioning skill. As a result both classroom
performance and written criterion measures were developed
for use in assessing treatment effects and ability-
performance relationships.
Classroom. Performance Measures
Several criterion measures were obtained from
transcripts of the classroom interaction of intern and
microstudents over three teaching sessions.
First, treatments were designed to produce in-
creases in Analytic Questioning as a general questioning
strategy. While different levels of significance might
be examined in relation to each subcategory of Analytic
Questioning_, differences of this nature were not emphasized
in the model presentation. Therefore, the total number of
analytic questions was selected as one appropriate index
for measuring increases in Analytic Questioning.
In addition to the total frequency, the variety
of categories of Analytic Questioning used was also of con-
siderable interest. Theoretically, the use of any category
of Analytic Questioning by Ss :represents an available class
of responses in the behavioral repertoire of that S as
well as a single occurrence of a behavior of interest.
Accordingly, for the purpose of this study, a total fre-
quency score representing the use of all categories would
demonstrate more desirable training effects than would
the same total frequency score representing the use of
fewer categories of Analytic Questioning.
Finally, while attention so far has been focused
exclusively on frequency measures, it was recognized that
the quality of the questions might also be considered a
potentially important dimension of the dependent variable.
Some analysis of quality had already been done in specify-
ing those kinds of statements fitting into the category
system. These categories were meant to include the best
examples of the behavior being learned. While weaker or
more ambiguous examples, which would not clearly meet the
criteria imposed by the category definition, would undoubt-
edly represent levels differing in quality, a considerable
degree of precision in classification was believed necessary
so that the extent of imitation could be assessed. However,
in developing the rating manual additional differences of
such a character were considered.
49
Quality is commonly defined as an approximation
to an ideal or standard. Accordingly, for the purposes of
this study, "quality" was used to distinguish between near
versus more remote approximations to the behavior exhibited
by the model. It had been observed in a previous study
(Koran, 1967) using a similar questioning strategy, that
one dimension along which questions in any given category
tended to vary was that of information content. Thus, in-
formation content was selected as the dimension along which
approximation to the behavior of the model would be assessed.
In all cases, the questioning behavior exhibited by the
model in this study required the student to supply the maxi-
mum amount of information that could be elicited by that
particular type of question. The closest approximations
to the model's behavior would also do this. In more remote
approximations, the teacher would supply varying amounts
of information which the model had elicited from the stu-
dents. Accordingly, questions meeting the criterion es-
tablished for high quality were designed to elicit from
the students all information relevant to answering them.
Questions requiring students to supply only part of the
information relevant to answering them, or requiring only
that students agree, disagree or select from among given
alternatives would not meet the criteria imposed by this
specification of quality.
To illustrate: the information content of a
semantic definition has been defined as a statement of the
methods and standards used for defining a word. The ques-
tion "How does the author define social class"? requires
students to supply the key elements of the definition as
well as the method of measurement, and would accordingly
be classified as a question of high quality. However, the
questions "How does the author define social class with
respect to income level"? and "Does the author define social
class in terms of occupation, education or income ? provide
part or all of the specified information content, and thus
would not meet the criteria established for high quality.
A complete discussion of quality, as used here, appears
in the Rater Manual ('ppendix A).
It should be noted that quality, as defined here,
is not intended to represent an evaluation of which kinds
of questions are "best." Any question could possibly be
made "better" or "worse" by changes which do not involve
approximation to the behavior of the model along the dimen-
sion of information, content. Attention is focused on the
quality dimension used here only to the extent that it can
be used to assess the extent of imitation.
51
Written Measures
While ultimately, decisions in teacher training
must be guided by the effectiveness of training procedures
in the transmission or modification of actual classroom
teaching behavior, written measures were introduced in an
effort to gather additional information useful in assessing
ability-performance relationships in observational learning?
These measures consisted of two recognition type tests
(Appendix D). These written tests were designed in the
following manner:
1. A true-false test in which subjects were asked
to identify the major categories of Analytic Ques-
tioning. (2 minutes)
2. A multiple choice test in which subjects were
asked to match a number of questions according
to membership in a particular subcategory of
Analytic Questioning. (3 minutes)
A common observation has been that differences
exist between measures of knowledge or comprehension, and
measures of actual performance in a given situation (Cron -
bath, 1963). In terms of the model proposed for investi-
gating individual differemces in learning that was pre-
viously discussed, ability-performance relationships may
52
be expected to vary as a function of the nature of the
task. While certain abilities, such as those associated
with verbal production, may be more decisively involved in
classroom performance measures, it is conceivable that dif-
ferent abilities, such as those related to perception and
memory, may be more likely to be involved in recognition
measures of the type used here.
Methods of Gathering Data
During the study each of the intern's lessons
was recorded on stenorette tape. Technicians collected
the tapes at the end of each lesson and labeled them with
each Ss n am e , treatment group and number of teaching ses-
sion. Each tape was monitored in order to assure that the
volume and length were adequate. Typists then prepared a
written transcript of each lesson. Each stenorette used
for transcribing was equipped with a timing device which
permitted the accurate recording of the ten minute com-
pletion time.
Written tests were collected following the third
teaching session. Each test was labeled with Ss name and
treatment group. Objective scoring keys were used for
both tests.
53
Procedures for Training Raters
Prior to the analysis of the experimental tapes,
a preliminary manual was developed from the written de-
scription of the dependent variable and analysis of the
model lesson. A team of four raters was given intensive
training on the definition of each category of Analytic
Questioning and the use of the preliminary rating manual.
Raters were trained using three complete tapes
(nine teaching sessions) representing the teaching sessions
of one S from each of the three treatment groups. In addi-
tion to these, the written model along with discarded ver-
sions of the film-mediated model were also used for train-
ing purposes. The experimenter and team of raters system-
atically rated each of the training scripts and tapes.
On the basis of experience with these, the preliminary
manual was revised, and a final Rater Manual was developed
(Appendix A). Approximately 40 hours were spent in training.
Once an acceptable level of agreement was reached,
the raters independently rated the experimental scripts.
Reliability was maintained throughout the analysis by
scheduling occasional meetings where raters would review
relevant decision rules and training scripts. This was
done to insure that systematic rating biases would not
54
develop. Neither ratings done during training nor in sub-
sequent meetings were used in the statistical analysis of
results.
Rating Procedures
Scripts were typed from audio tapes of each teach-
ing session and were coded according to SI treatment and
number of teaching session. This was done to insure that
the raters had no knowledge of the Ss name, treatment con-
dition, or phase in the treatment condition. In each
script, all questions were subsequently marked and numbered.
Scripts were than subdivided into four groups and distri-
buted to raters. When scripts were completed by one judge
they were exchanged with the three other judges until each
judge had rated every script. Each rater used a standard
form (Appendix E) for recording the occurrence o° Analytic
Questioning behavior. Rater scores were subsequently
transferred to coding sheets and data cards for analysis.
Rater Reliability
The reliability .of scores was established for
judges 1, 2, 3, and 4 over all rated categories. The
55
Analysis of Variance repeated measures model described
by Winer (1962) was used for this purpose. This analysis
provides two reliability coefficients: (1) an estimate
of the reliability of the mean of four ratings, which
represents the expected correlation between mean ratings
for the same people with other random samples of judges,
and (2) an estimate of the reliability of a single mea-
surement, which is approximately equal to the average
intercorrelation between ratings given by pairs of judges.
The reliability of both the mean scores and of
a single measurement for each rated category for each
teaching session appears in Table 5. These coefficients
are extremely high, suggesting that the criterion measures
are sufficiently reliable for testing both treatment ef-
fects and ability-performance relationships.
The reliability coefficients reported are some-
what higher than expected on the basis of the general
literature on rating. The extensive pretraining of raters,
the specificity of the behavior rated, and the derivation
of the ratings from static transcripts rather than from
ongoing behavioral sequences are believed to have contri-
buted to the high reliability coefficients.
TABLE 5
RATER RELIABILITY
TeachingSession 1
TeachingSession 2
TeachingSession 3
Category a b a b a
Total Analytic .95 .88 .99 .96 .99 .97
Hypotheses .95 .87 .96 .88 .96 .88
Definitions .94 .85 .97 .90 .97 .94
Assumptions .98 .93 .99 .97 .99 .97
Facts .97-, .90 .97 .94 .97 .93
Conclusions .90 .75 .95 .87 .96 .95
Total Categories .96 .90 .98 .94 .98 .95
High Quality Level .92 .81 .98 .94 .98 .95
Total Non Analytic .98 .95 .99 .97 .99 .96
a - Figures in this column representmean scores.
coefficients of the reliability of
b - Figures in this column representsingle measurement.
coefficients of the reliability of a
CHAPTER III
RESULTS
The primary objectives of this study were:
1. To assess the relative effects of different
modes of model presentation.
2. To explore the effects of individual differ-
ences on observational learning.
This chapter will describe the statistical tests
of the hypotheses and the results achieved. The presen-
tation of results will treat first, the instructional
treatment main effects; second, the ability measures;
and finally, the aptitude x treatment interactions.
A complete list of all independent and dependent
measures to be analyzed appears in Table 6. The distribu-
tions of scores for the dependent measures permitted evalua-
tion by parametric techniques. Most analyses were computed
using UCLA Bio-Medical Programs. (Dixon, 1964)
57
TABLE 6
INDEPENDENT AND DEPENDENT MEASURES
58
Independent Measures Dependent Measures
Hidden Figures
Hidden Figures
Hidden Figures
Identical Pictures
Identical Pictures
Identical Pictures
Maze Tracing
Maze Tracing
Maze Tracing
Word Arrangement
Word Arrangement
Word Arrangement
- Part I
- Part II
- Total
- Part I
- Part II
- Total
- Part I
- Part II
- Total
- Part I
- Part II
- Total
Advanced Vocabulary - Part I
Advanced Vocabulary - Part II
Advanced Vocabulary - Total
Verbal GRE
Film Memory
Memory for Ideas -
Memory for Ideas -
Memory for Ideas -
Sentence Reproduction-
Sentence Reproduction-
Sentence Reproduction-
Part I
Part II
Total
Part I
Part II
Total
Analytic Questions, T1
Categories of Analytic Questions, T1
High Quality Analytic Questions, T1
Nonanalytic Questions, T1
Analytic Questions, T2
Categories of Analytic Questions, T2
High Quality Analytic Questions, T2
Nonanalytic Questions, T2
Analytic Questions, T3
Categories of Analytic Questions, T3
High Quality Analytic Questions, T3
Nonanalytic Questions, T3
True-False Test
Matching Test
59
Instructional Treatment Main Effects
Classroom Performance Measures
The means of classroom performance measures were
derived for each S by averaging the scores given on each
category by each of the four judges. All analyses were
computed using these mean scores.
The results for the four 'classroom performance
measures previously described were analyzed separately for
each of the three teaching sessions, although in some cases,
the relationship among these measures was sufficiently high
to suggest that they do not represent psychologically dif-
ferent variables. This was done in view of differences in
ability-performance relationships observed and because of
the unavailability of computer programs for multivariate
analysis of variance techniques. Intercorrelations among
dependent measures are reported both for the total sample
(Table 7) and for each experimental group separately (Ap-
pendix F).
The initial test used to determine if there were
significant instructional treatment main effects on the
criterion variables was a 3 x 3 repeated measures analysis
of variance. Since conformity of the data to the homogeneity
TABLE 7
INTERCORRELATIONS AMONG DEPENDENT VARIABLES FOR TOTAL SAMPLE
2 3 4 5 6 7 8 9 10 11 12 13 14
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Analytic Questions, T1
Categories, T1
High Quality, T1
Nonanalytic, T1
Analytic Questions, T2
Categories, T2
High Quality, T2
Nonanalytic, T2
Analytic Questions, T3
Categories, T3
High Quality, T3
Nonanalytic, T3
True-False Test
Matching Test
77 83
68
15
19
18
29
14
20
03
23
15
14
-07
78
29
14
22
01
95
74
-02 29
06 14
-04 23
34 09
-30 74
-36 64
-31 68
-30
22
10
14
-06
70
73
67
-31
83
25 09
10 15
25 11
06 44
73 -2.8
62 -2S
71 -29
033 65
94 -27
81 -27
-31
08
00
-05
06
49
40
45
-21
48
46
48
-24
-07
-07
-16
-16
42
43
42
-21
40
47
42
-29
56
Decimals omitted
61
assumptions underlying the use of this model appeared ques-
tionable (Appendix G), separate one and two-way analyses
of variance were also computed for each measure of the
dependent variable. Since the results were the same for
both methods of analysis, the repeated measures analysis
of variance was selected as the more parsimonious way of
reporting these data. However, the results of the separate
one and two-way analyses of variance are reported in Ap-
pendix H.
In accordance with Winer (1962), when the inter-
action between the two factors in the repeated measures
analysis of variance was significant, tests on simple main
effects were computed in addition to direct tests on main
effects. In such cases, both direct tests on main effects
and tests on simple main effects are reported separately.
In all cases, the Newman-Keuls procedure (Winer, 1962) was
used in comparisons of pairs of treatments following a
significant overall F ratio (Appendix I).
Table 8 presents the means and standard deviations
of the classroom performance measures for each of the three
experimental groups.
Analytic Questions. The results for the repeated
measures analysis of variance of the frequency of Analytic
TABLE 8
MEAM_AND STANDARD DEVIATIONS OF DEPENDENT. VARIABLES
Performance Measure
Treatment GroupFilm - Mediated
ModelingWrittenModeling Control
Mean S.D. Mean S.D. Mean S.D.
Analytic Questions, T1 3.61 3.37 3.62 3.30 3.72 3.10Categories of Analytic Questions, T1 1.43 1.12 1.75 1.01 1.76 1.11High Quality Analytic Questions, T1 1.88 2.02 1.75 1.89 2.18 2.24Nonanalytic Questions, T1 23.26 9.41 24.39 9.33 25.24 10.80Analytic Questions, T2 13.40 6.41 8.16 5.59 3.31 2.50Categories of Analytic Questions, T2 3.73 1.19 2.89 1.39 1.59 1.05High Quality Analytic Questf.ons, To 9.90 5.20 5.17 3.84 1.92 1.70Nonanalytic Question, T2 ` 16.02 9.36 19.12 11.49 25.40 10.57Analytic Questions, T3 13.58 6.33 10.09 7.11 2.93 3.25Categories of Analytic Questions, T3 3.85 1.17 2.98 1.37 1.33 .97High Quality Analytic Questions, T3 10.15 5.52 6.72 4.42 1.68 2.24Nonanalytic Questions, T3 15.50 8.16 17.23 9.15 25.96 12.72Matching Test 11.82 2.36 10.67 3.09 7.97 2.88True-False Test 10.70 2.02 9.77 2.35 7.00 1.90
63
Questions are presented in Table 9. In view of the signif-
icant interaction between treatment groups and teaching
sessions, (F = 29.41, p < .01) tests on simple main effects
were computed (Table 10). These tests disclosed highly
significant treatment effects both for T2 (F = 46.27, p<.01),
and for T3 (F = 50.55; p < .01), while no differences be-
tween groups were found for T1 (F = .01). Comparisons be-
tween pairs of treatment conditions using the Newman-Keuls
procedure showed that for both T2
and T3, the Written and
Film-Mediated Modeling groups generated significantly more
Analytic Questions (p < .01) than did the Control group.
Moreover, Ss in the Film-Mediated Modeling condition pro-
duced significantly more Analytic Questions than did Ss
in the Written Modeling condition on both T2 and T3
(p < .01).
An alternative view of the relative efficacy of
the different instructional treatments in augmenting the
frequency of Analytic Questions is furnished by within-
group analyses of the increase from base rate to subsequent
teaching sessions. The increase was found to be significant
for both the Film-Mediated (F = 30.13; p < .01) and Written
(F = 9.74; p < .0.1) Modeling groups. The Control group did
not display increments in the frequency of Analytic Ques-
tions (F = .14). These modeling effects are illustrated in
Figure 2.
64
TABLE 9
ANALYSIS OF VARIANCE OF ANALYTIC QUESTIONS
Source of Variation df MS
Between Subjects 119
Treatment Groups 2 1422.66 32.10 .01
Subjects Within Grow. 44.32
Within Subjects 240
Teaching Sessions 2 992.67 75.31 .01
Treatments X Sessions 4 387.64 29.41 .01
Sessions x Subjects Within Groups 234 13.18
65
TABLE 10
SUMMARY OF ANALYSIS OF VARIANCE OF ANALYTIC QUESTIONSSIMPLE MAIN EFFECTS
Source of Variation cif MS
Beteen Groups
Treatments at Teaching Session 1 2 .16 .01 NS
Treatments at Teaching Session 2 2 1030.25 46.27 .01
Treatments at Teaching Session 3 2 1191.01 50.55 .01
Error Between Groups 351 23.56
Within Groups
Sessions of Film-Mediated Modeling 2 1335.51 30.13 .01
Sessions of Written Modeling 2 442.00 9.74 .01
Sessions of Control 2 6.30 .14 NS
Error Within Groups 234 13.18
.... mot
Amwwwwww====m-,47
14
12
I0
66
MOM
MID
FILM -MEDIATEDMODELING
WRITTEN MODELING
I
TI
CONTROL
T2 T3
FIGURE 2. MEAN FREQUENCY OF ANALYTIC QUES-TIONS ACROSS TEACHING SESSIONS.
67
Categories of Anal tic Questions. Table 1.11' pre-
sents the results for the repeated measures analysis of
variance for the variety of categories of Analytic Questions
used. Since there was a significant interaction (F = 27.65,
p <.01) between treatment groups and teaching sessions,
tests for simple main effects were also performed (Table 12).
These analyses yielded significant treatment effects both
for T2 (F = 34.23; p <.01) and for T3 (F = 48 .21; p <.01).
Differences between treatment groups were not found for
T1 (F = .97). The specific differences contributing to
the treatment effects were therefore investigated by com-
parisons of pairs of treatments. The comparisons showed
that both the Written and Film-Mediated Modeling groups
used significantly more categories of Analytic Questions
both for T2
and T3
(p < .0 ) than did the Control group.
In addition, Ss in the Film-Mediated Modeling condition
used significantly more categories of Analytic Questions
(p < 01) than did Ss in the Written Modeling condition
for both teaching sessions.
Within-group analyses of change for each
group between T1 and T3 showed a significant increase in
the number of categories of Analytic Questions used both
for the Written Modeling group (F = 21.81; p.01) and
68
TABLE 11
ANALYSIS OF VARIANCE OF CATEGORIES OF ANALYTIC QUESTIONS
Source of Variation df MS F p
Between Subjects 119
Treatment Groups 2 64.79 24.75 .01
Subjects Within Groups 117 2.36
Within. Subjects 240
Teaching Sessions 2 47.14 54.17 .01
Treatments X Sessions 4 24.06 27.65 .01
Sessions X Subjects Within Groups 234 .87
TABLE 12
SUMMARY OF ANALYSIS OF VARIANCE OF CATEGORIES OF ANALYTIC QUESTIONS
SIMPLE MAIN EFFECTS
69
Source of Variation df MS
Between Groups
Treatments at Teaching Session 1 2 1.34 .97 NS
Treatments at Teaching Session 2 2 46.90 34.23 .01
Treatments at Teaching Session 3 2 66.05 48.21 .01
Error Between Groups 351 1.37
Within Groups
Sessions of Film-Mediated Modeling 2 75.65 86.95 .01
Sessions of Written Modeling 2 18.98 21.81 .01
Sessions of Control 2 1.84 2.11 NS
Error Within Groups 234 .87
70
FILM-MEDIATED MODELIN
WRITTEN MODELING
CONTROL
T3
FIGURE 3. MEAN FREQUENCY OF CATEGORIES OF
ANALYTIC QUESTIONS ACROSS TEACH-
ING SESSIONS.
71
for the Film-Mediated group (F = 86.95; p <. 01), while
the Control group did not show increases in the number
of categories used (F = 2.11).
Figure 3 depicts the mean number of categories
used by each group for each teaching session.
High Quality Analytic Questions. Results from
the repeated measures analysis of variance of the quality
of Analytic Questions appear in Table 13. Testri for simple
main effects were performed (Table 14) in view of the
significant interaction (F = 32.61: p<. 01) between treat-
ment groups and teaching sessions. These tests disclosed
highly significant treatment effects for T2 (F = 51.37;
p<.01) and for T3 (F = 57.80; p<.01), while treatment
effects were not found for T1 (F = .08). Comparisons
b'itween pairs of treatment conditions showed that Ss in
both the Written and Film-Mediated Modeling groups generated
significantly more high quality Analytic Questions than
did Ss in the Control group, both for T2 and for T3 (p <.01).
Similarly, the two modeling procedures proved to be dif-
ferentially effective, with Ss in the Film-Mediated Modeling
group producing significantly more high quality Analytic
Questions than did Ss in the Written Modeling group (p<.01)
for both teaching sessions.
ANALYSIS OF VARIANCE OF HIGH QUALITY ANALYTIC QUESTIONS
Source of Variation
72
F p
Between Subjects
Treatment Groups
Subjects Within Groups
Within Sublects
Teaching Sessions
Treatments X Sessions
Sessions X Subjects Within Groups
119
2 874.83 38.67 .01
117 22.62
240
2 614.88 79.74 .01
4 251.48 32.61 .01
234 7.72
TABLE 14
SUMMARY OF ANALYSIS OF VARIANCE OF HIGH QUALITY ANALYTIC QUESTIONS
SIMPLE MAIN EFFECTS
Source of Variation df MS F p
Between Groups
Treatmentsat Teaching Session 1 2 .90 .08 NS
Treatmentsat Teaching Session 2 2 652.46 51.37. .01
Treatmentsat Teaching Session 3 2 734.11 57.80 .01
Error Between Groups 351 12.69
Within Groups
Sessions of Film-Mediated Modeling 2 906.27 117.39 .01
Sessions of Written Modeling 2 225.52 29.25 .01
Sessions of Control 2 1.33 .32 NS
Error Within Groups 234 7.72
IO
8
6
4
2
0
FILM-MEDIATED MODELING
WRITTEN MODELING
CONTROL
I I I
74
Ti T2 T3
FIGURE 4. MEAN FREQUENCY OF HIGH QUALITYANALYTIC QUESTIONS ACROSS TEACH-ING SESSIONS.
75
Within-group analyses provided another view of
the differential effectiveness of the instructional treat-
ments in producing high quality Analytic Questions. Ss in
both the Film-Mediated Modeling group (F = 117.39; p < .01)
and Written Modeling group (F = 29.25; p < .01) displayed
significant increases in the number of high quality Analy-
tic Questions while Control group Ss did not display signif-
icant increments (F = .32). A visual examination of these
modeling effects appears in Figure 4.
Nonanalytic Questions. It was anticipated that
treatment effects might manifest themselves by a rise in
the frequency of Analytic Questions in some groups and a
related decrease in the frequency of Nonanalytic Questions.
Table 15 presents the results for the repeated measures
analysis of variance of Nonanalytic Questions. Since a
significant interaction between treatment groups and teach-
ing sessions was disclosed (F = 4.47; p < .05), tests for
simple main effects were also computed (Table 16). These
analyses showed significant treatment effects both for T2
(F = 8.04; p < .01) and for T3 (F = 12.08; p < .01). Sig-
nificant treatment effects were not found for T1 (F = .38).
Differences contributing to the treatment effects were in-
vestigated by comparing pairs of treatments. The comparisons
76
TABLE 15
ANALYSIS OF VARIANCE OF NONANALYTIC QUESTIONS
Source of Variation df MS
Between Subjects 119
Treatment Groups 2 1719.19 8.56 .01
Subjects Within Groups 117 200.59
Within Subjects 240
Teaching Sessions 2 792.34 14.02 .01
Treatments X Sessions 4 252.68 4.47 .05
Sessions X Subjects Within Groups 234 56.48
77
TABLE 16
SUMMARY OF ANALYSIS OF VARIANCE OF NONANALYTIC QUESTIONSSIMPLE MAIN EFFECTS
Source of Variation df MS
Between Groups
Treatment at Teaching Session 1 2 39.73 .38 NS
Treatment at Teaching Session 2 2 921.11 8.04 .01
Treatment at Teaching Session 3 2 1262.89 12.08 .01
Error Between Groups 351 104.51
Within Groups
Film-Mediated Modeling 2 772.13 13.67 .01
Written Modeling 2 550.23 9.74 .01
Control 2 5.60 .10 NS
Error Within Groups 234 56.48
78
26
24
22
20
18
16
14TI T2 13
FIGURE 5. MEAN FREQUENCY OF NONANALY-TIC QUESTIONS ACROSS TEACH-ING SESSIONS.
79
showed that for T2 and T3, Control group Ss produced sig-
nificantly more Nonanalytic Questions than did Ss in either
the Written or Film-Mediated Modeling conditions (p < .01).
However, the Written and Film-Mediated Modeling groups did
not differ significantly from one another on either occasion.
Differential treatment effects are also shown
by within-group analyses of changes in performance from
T1 to T3. Is in the Film-Mediated Modeling group (F = 13.67;
p < .01) and the Written Modeling group (F = 9.74; p < .01)
displayed significant decreases in the number of Nonanaly-
tic Questions used, while Control group Ss did not display
such decrements (F = .10). It should also be recalled
here that Ss in the Written and Film-Mediated Modeling
conditions displayed significant increases in the frequency
of Analytic Questions, while Control group Ss did not.
These modeling effects are illustrated in Figure 5.
Written Measures
As previously indicated, performance on two
written measures was obtained in addition to classroom
performance measures. The written measures consisted of
a True-False Test in which Ss were asked to identify examples
80
and nonexamples of major categories of Analytic Question-
ing; and a Matching Test, in which subjects were asked
to match correctly given questions with categories of
Analytic Questions.
A one-way analysis of variance was computed be-
tween groups to determine if there were significant in-
structional treatment main effects for the written mea-
sures. Following significant F ratios, the Newman-Keuls
procedure was used in comparisons of specific pairs of
treatments (Appendix (T). Table d presents the means and
standard deviations of the written measures for each of
the three experimental groups.
True-False Test. The results for analysis be-
tween treatment groups of scores on the True-False Test
are shown in Table 17. A one-way analysis of variance
computed between the three groups disclosed significant
treatment effects (F = 20.76; p < .01).
Ss in the Written and Film-Mediated Modeling
conditions correctly identified significantly more examples
and nonexamples of categories of Analytic Questions than
did Ss in the Control group (p < .01). However, scores
for Ss in the Written and Film-Mediated Modeling condi-
tions did not differ significantly from one another.
TABLE 17
ANALYSIS OF VARIANCE OF SCORES FOR TRUE-FALSE TEST
Source of Variation df MS P
Between Croups 2 162.24 20.76 .01
Within Groups 117 7.81
..111
81
Matching Test Table 18 presents the results
for analysis between treatment groups of scores for the
Matching Test. This analysis showed highly significant
treatment effects (F = 32.94; p< .01).
Comparisons between pairs of treatment conditions
showed that Ss in the Written and Film-Mediated Modeling
conditions correctly matched significantly more items on
the test than did Ss in the Control group (p < .01). More-
over, Ss in the Film-Mediated Modeling conditioti obtained
significantly higher scores (p < <05) than did Ss in the
Written Modeling condition.
82
TABLE 18
ANALYSIS OF VARIANCE OF SCORES FOR MATCHING TEST
Source of Variation df MS F P
Between groups
Within Groups
2 145.38 32.94 .01
117 4.41
Ability Measures
Scores on a series of nine ability measures were
obtained for 120 Ss included in the analysis. Following the
completion of testing, Ss were randomly assigned to the three
experimental groups. Table 19 presents the means and stan-
dard deviations of these ability measures for the total sample
in addition to the composition of the three groups separately.
Frequency distributions of the ability measures
were judged to be approxinately normal. Thus, normality was
assumed for the purposes of further analysis.
Reliability
Reliabilities for all tests used, with the excep-
tions of Film Memory and the Graduate RecordsExamination,
TABLE 19
MEANS AND STANDARD DEVIATIONS OF ABILITY MEASURES
Treatment Group
Film-Mediated WrittenModeling Modeling Contr )1
TotalSample
Mean S.D. Mean S.D. Mean S.D. Mean S.D.
Hidden Figures - Part IHidden Figures - Part IIHidden Figures - TotalIdentical Pictures - Part IIdentical Pictures - Part IIIdentical Pictures - TotalMaze Tracing - Part IMaze Tracing - Part IIMaze Tracing - TotalWord Arrangement - Part IWord Arrangement - Part IIWord Arrangement - TotalVerbal Comprehension - Part IVerbal Comprehension - Part IIVerbal Comprehension - TotalVerbal G.R.E.Film MemoryMemory for Ideas - Part IMemory for Ideas - Part IIMemory for Ideas - TotalSentence Reproduction - Part ISentence Reproduction - Part IISentence Reproduction - Total
6.34 4.00 6.47 3.46 6.10 3.92 6.31
6.14 3157 6.87 3.; 63 6.21 3..59 6.4112.48 6.75 13.35 6..17 12.31 7.01 12.741.26 8.09 41.70 5.50 40.68 6.54 41.2339.83 8.13 41.42 6.56 39.47 6.80 40.2580.71 14.90 82.62 12.05 79.95 12.27 81.1110.10 3.45 11.15 3.18 10.86 4.10 10.7013.46 4.07 14.13 3.77 14.63 4.25 14.0623.56 6.98 25.27 6.46 25.55 7.94 24.7719.82 6.71 20.97 6.62 20.18 7.16 20.3322.99 7.83 24.30 7.83 21.11 7.54 22.5541.78 13.63 45.02 13.80 40.89 13.83 42.5913.68 2.92 14.10 2.58 14.03 2.44 13.9314.34 3.19 14.35 2.51 14.21 3.07 14.3027.98 5.76 28.48 4.71 28.23 5.18 28.20
611.95 97.06 582.75 158.00 582.36 139.71 592.6920.34 3.96 20.80 4.70 19.97 3.12 20.3823.00 6.57 22.90 7.87 22.47 6.04 22.8014.93 5.11 16.05 5.71 13.82 4.35 14.9537.85 10.80 38.45 10.20 36.29 9.06 37,55136.93 17.83 140.85 15.50 137.65 12.49 138.48139.29 26.69 143.07 30.56 138.23 18.27 140.23276.22 41.01 281.89 24.27 275.89 26.17 278.03
3.773.586.616.767.20
13.103.594.037.136.797.78
13.752.652.915.19.
133.273.986.83,5.1410.0315.4625.6734.36
84
were estimated from a product-moment correlation between
scores on two separately timed parallel parts of each test.
The Spearman-Brown formula was used to obtain the reliabil-
ity coefficients for the full tests.
KR20 estimates of realiability for the Graduate
Records Examination were available in the manual provided
by test publishers. Reliability of Film Memory was also
computed using the Kuder-Richardson formula.
Reliabilities for all tests are provided in
Table 20, along with a standard error of measurement com-
puted for each ability measure. Reliabilities of factor
tests from the French Kit were consistently higher than
those of other factor tests used. Reliabilities for the
audio-visual tests were acceptable, although somewhat lower.
Intercorrelations ofAbility Measures
Intercorrelations of ability measures were com-
puted both for the total sample (Table 21) and for each
of the three experimental groups separately (Appendix J).
For the most part, the factor tests display low
correlations with one another. This is in accordance with
85
TABLE 20
RELIABILITIES AND STANDARD ERRORS OF MEASUREMENTFOR ABILITY MEASURES
Test Reliability
Standard Errorof Measurement
Hidden Figures (Cf-1) .76 3.23
Identical Pictures (P-3) .83 5.37
Maze Tracing (Ss-1) .86 2.70
Word Arrangement (Fe-3) .82 5.77
Advanced Vocabulary (V-4) .87 1.86
Memory for Ideas .65 5.91
Sentence Reproduction .69 19.24
Graduate Records Examination(Verbal Score) .92 28.0u
Film Memory .50 2.98...
TABLE 21
INTERCORRELATIONS OF ABILITY MEASURES:
TOTAL SAMPLE
12
34
56
78
910
11
12
13
14
15
16
17
16
19
20
21
22
23
1.
2.3,
4.
5.
6.
7.8.9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
Hidden Figures - Part I
Hidden Figures - Part II
Hidden Figures - Total
Identical Pictures - Part I
Identical Pictures - Part II
Identical Pictures - Total
Maze Tracing - Part I
Maze Tracing - Part II
Maze Tracing - Total
Word Arrangement - Part I
Word Arrangement - Part II
Word Arrangement - Total
Verbal Comprehension - Part I
Verbal Comprehension - Part II
Verbal Comprehension - Total
Verbal GRE
Film Memory
Memory For Ideas - Part I
Memory For Ideas - Part II
Memory For Ideas - Total
Sentence Reproduction - Part I
Sentence Reproduction - Part II
Sentence Reproduction - Total
61
9089
34
3337
33 42
4171
34
4142
88 93
3125
32384345
2829
31
434549
75
32
2934
4347
50
9294
2525
28
210815
071612
1014
131820
221019
1668
1720
2120
15
2010
20
1689
93
06 10
09
27 25
2909
1714
1320
18
15
1316
32
3335
2019
211019
1777
12
13
13
31
3134
161919
122019
9394
150712
2724
3030
2328
13161748
5756
-04
1305
-04
06
021609
1314
0610
-02060203
0815
1220
1217
13
04
0924
21 23
070909
08
27
-020702
23182019
12
1614
20
18
11111210
33 48
04
1309
25
16
2117
10 1424
2526
1112
1310
37 90
75
0101
01 1609
11
08
0205
08
03072432
2925
02
28
25
31
0608
0719
131610
05
08
0709
101930
27
30
0131 15
28
53
0205
0419
1215
09
0205
08
0609
2536
3332
02
3422
3384
87
Decimals omitted.
87
the expectations regarding correlations among factor tests
presumed to be independent. However) those factor tests
from the French Kit which may be regarded as measures of
perceptual abilities (Hidden Figures, Identical Pictures,
Maze Tracing) tend to correlate more highly with one another
than they do with other tests. These correlations, while
low, are consistently significant. Measures of verbal
performance display a somewhat weaker tendency to cor-
relate more highly with one another than with primarily
perceptual measures.
Relation of Ability Measuresto Performance
The major purpose of this study, it will be re-
called, was to examine the effects of individual differences
on observational learning. In terms of the model previously
proposed by Melton (1967) for investigating individual dif-
ferences in learning, ability-performance relationships
may be expected to vary as a function of the nature of
the task. It was anticipated that the requirements of
the Written and Film-Mediated Modeling conditions were
sufficiently different to produce different ability-
performance relationships.
88
A first step in evaluating these theoretically
expected relationships was to compute for each group sepa-
rately, the correlations between all ability and all per-
formance measures. Both part scores and total scores were
used for this purpose. These correlations appear in Tables
22, 23, and 24. Upon inspection it was apparent that in
several instances the correlations between ability and
performance measures varied substantially across treatment
conditions. The most promising variables (i.e., those
variables which demonstrated the greatest variation among
the groups) were then selected for the investigation of
aptitude X treatment interactions. Only variables which
showed correlations significantly different from zero were
used in this analysis.
It should be noted that in several instances,
there are substantial differences between the correlations
of part scores of a single ability measure with a given
performance measure. The most notable example of this
can be seen in the part scores for Hidden Figures, where
it is obvious that Part I and Part II scores are predict-
ing differentially. Since ultimately we are interested
in predicting learning, in cases such as these, part scores
rather than total scores have been used in subsequent
analysis.
TABLE 22
CORRELATIONS OF ABILITY MEASURES WITH PERFORMANCE MEASURES
FILM-MEDIATED MODELING GROUP
T1
T2
Hidden Figures - Part I 03 08 02 20 -15 -20
Hidden Figures - Part II -13 -14 -11 17 10 -14
Hidden Figures - Total -05 -02 -05 21 -04 -19
Identical Pictures - Part I 22 35 13 14 -13 -11
Identical Pictures - Part II 34 34 32 11 13 -02
Identical Pictures - Total 31 36 24 17 03 -06
Maze Tracing - Part I 00 18 03 07 -10 -09
Maze Tracing - Part II 11 21 25 18 -12 -25
Maze Tracing - Total 06 21 16 14 -12 -19
Word Arrangement - Part I 19 22 12 37 -05 -21
Word Arrangement - Part II 21 21 17 40 29 -09
Word Arrangement - Total 23 26 17 44 14 -17
Advanced Vocabulary - Part I 28 34 11 33 21 00
Advanced Vocabulary - Part II 24 30 07 30 20 03
Advanced Vocabulary - Total 27 34 10 34 20 01
Verbal GRE 36 43 16 15 17 -05
Film Memory -02 01 -01 04 09 01
Memory for Ideas - Part I 27 19 09 09 17 21
Memory for Ideas - Part II 41 32 35 07 17 08
Memory for Ideas - Total 41 36 26 11 18 17
Sentence Reproduction - Part I 30 14 15 21 22 00
Sentence Reproduction - Part II 26 11 13 29 38 19
Sentence Reproduction - Total 30 13 15 28 34 12
Decimals Omitted
r = .26, p < .05
r = .36, p < .01
T34)
CD
E-1
04)Cif
rob
4-1
.0ttO4-1
4-1
fI
00
04-1
Ha0
l
1:C
4-1Cif0
4--1
.0tIO
4-1
04--1
H0Cif
cts
00
-20 05 -25 -49 -20 00 11
00 18 15 -19 06 -04 16
-12 13 -07 -39 -09 -01 15
-09 27 -13 -10 -14 16 03
13 04 08 05 06 02 04
06 08 02 02 00 10 04
-18 21 -13 -14 -21 lg -22
-16 23 -07 -13 -08 28 -13
-18 24 -10 -15 -15 25 -18
-13 40 04 -06 01 56 13
26 39 21 13 21 47 26
08 45 15 03 13 56 21
11 45 12 14 -01 38 23
11 39 05 02 -01 32 33
10 43 09 07 -02 36 31
07 22 -01 -07 -08 27 43
05 -13 39 34 .44 -03 05
11 11 06 -10 03 02 07
17 20 10 12 14 07 10
15 18 07 -02 07 03 11
11 12 -11 -23 -17 16 28
34 27 14 -09 09 14 19
27 23 04 -16 -02 17 24
89
16
30
25
18
30
28
01
00
00
12
30
23
11
28
21
28
-25
23
06
19
-05
21
11
90
TABLE 23
CORRELATIONS OF ABILITY MEASURES WITH PERFORMANCE MEASURES
WRITTEN MODELING GROUP
T1
T2
T3Da
.0 0 .00 H H 0 T 4
AlP1-44-4t O2 t O
c 4.) 00 Z at 4.) ' 40Z as ri 0 0 as
G4 c..) = aC 0 =
0)4.) 0
t0 .g..4 HW H
g-4 $.10H9-4tliO CD' 2CZ 4.) 00 Zg as r-1 0
G4 0 = M
.0 W0 E-
0 WW4 E
Al0HZas2 9-I gil4g 4.) g
al AlE E4
Hidden Figures - Part I 28 34 18 14 38 29 36 -05 46 35 45 01 30 -20
Hidden Figures - Part II -08 11 -09 -02 09 01 05 11 -08 -07 -09 02 12 -02
Hidden Figures - Total 11 26 05 07 26 16 23 03 21 15 20 02 24 -14
Identical Pictures - Part I -02 14 -01 10 C8 08 08 -04 13 -06 16 -04 31 -05
Identical Pictures - Part II -02 21 00 18 03 04 03 18 07 -03 08 07 33 03
Identical Pictures - Total 01 20 02 13 03 02 06 11 05 -08 07 06 34 00
Maze Tracing - Part I 05 01 -01 -30 -02 -09 06 02 10 05 11 -20 25 19
Maze Tracing - Part II 05 12 -1- e3 15 06 19 01 20 24 26 -02 55 41
Maze Tracing - Total 05 07 -07 -28 07 01 14 01 17 16 21 -11 45 34
Word Arrangement - Part I 06 22 -03 06 07 12 03 25 18 25 16 19 09 05
Word Arrangement - Part II -02 13 -04 00 -02 04 -02 14 06 20 10 07 02 02
Word Arrangement - Total -01 16 -07 -02 -01 07 -02 21 07 22 10 12 00 04
Advanced Vocabulary - Part I 19 37 13 26 13 -02 18 00 -01 -19 -09 21 08 -09
Advanced Vocabulary - Part II 21 37 11 13 20 08 19 -02 21 -02 10 09 12 00
Advanced Vocabulary - Total 23 37 14 18 14 01 18 -03 08 -16 -01 15 09 -05
Verbal GRE 23 10 15 -01 -11 -06 -07 08 13 -02 01 06 09 -12
Film Memory -03 08 01 12 08 23 04 20 03 14 -02 02 22 32
Memory for Ideas - Part I 14 18 18 15 17 15 18 -19 07 09 07 -21 08 -24
Memory for Ideas - Part II 15 02 12 22 25 25 22 -08 28 22 23 -22 12 07
Memory for Ideas - Total 12 10 16 20 22 21 22 -17 17 19 15 -24 20 -05
Sentence Reproduction - Part I 15 26 19 -07 29 27 31 -13 19 14 20 -28 25 25
Sentence Reproduction - Part II 24 10 09 -30 14 14 27 -25 13 12 23 -34 22 08
Sentence Reproduction - Total 22 21 18 -24 21 20 29 -23 14 10 21 -38 23 16
Decimals emitted.
r= .26, p< .05
r = .36, p< .01
TABLE 24
CORRELATIONS OF ABILITY MEASURES WITH PERFORMANCE MEASURES
CONTROL GROUP
T T2
434-1
cd0
Hidden Figures - Part I 17 07 21 01 -04 -14 09 -14 05 03 -04 09 -15 -02
Hidden Figures - Part II 07 01 02 -09 -18 09 -04 00 -02 -02 03 -11 03
Hidden Figures - Total 13 04 11 01 -07 -17 10 -10 03 00 -04 07 -14 03
Identical Pictures - Part I 18 26 19 16 15 09 31 34 18 19 13 29 14 -09
Identical Pictures - Part II 20 22 16 09 09 04 26 22 13 17 17 29 -06 03
Identical Pictures - Total 19 23 18 14 14 06 32 29 17 18 16 31 04 -06
Maze Tracing - Part I 22 28 24 02 30 32 44 02 37 52 33 30 -15 -11
Maze Tracing - Part II 14 21 11 03 12 11 25 17 28 37 28 19 -11 -24
Maze Tracing - Total 18 25 18 02 22 23 36 11 34 47 32 26 -13 -18
Word Arrangement - Part I 05 04 -04 -03 -11 -18 -09 15 09 -01 04 12 07 26
Word Arrangement - Part II 18 -01 17 14 -13 -21 -05 24 21 05 19 29 -01 20
Word Arrangement - Total 15 04 10 07 -11 -22 -07 20 18 04 14 24 03 22
Advanced Vocabulary - Part I 13 -02 11 -30 33 28 20 -23 17 12 09 -05 -25 -05
Advanced Vocabulary - Part II 22 12 09 -29 49 29 38 -19 09 15 -02 06 -05 08
Advanced Vocabulary - Total 19 07 11 -32 45 30 32 -23 13 15 03 01 -15 02
Verbal GRE 20 10 05 -13 37 23 34 -01 14 27 07 03 02 08
Film Memory 23 17 -02 -27 24 22 28 -05 26 39 18 18 18 30
Memory for Ideas - Part I 20 31 21 -04 12 03 09 -02 35 26 33 04 19 07
Memory for Ideas - Part II 04 05 -06 09 -10 -08 -07 07 16 24 17 05 13 05
Memory for Ideas - Total 15 23 11 01 03 -01 03 02 30 29 30 05 19 07
Sentence Reproduction - Part I -05 -12 -02 -26 -11 -22 -07 -25 04 01 -03 -07 -03 16
Sentence Reproduction - Part II -01 -10 -13 -14 19 05 -13 -07 12 06 05 -12 14 -05
Sentence Reproduction - Total -03 -13 -09 -22 07 -07 06 -17 10 05 02 -12 08 04
Decimals omitted.
r = .26, p< .05
r = .36, p< .01
92
One possible explanation for the differential
functioning of the part scores might center on the nature
of the particular tests being considered. These tests
are both unique in nature and high in difficulty level.
Accordingly, the two parts of the tests can be thought
of as two separate trials in a learning task. It
is conceivable that differences in part scores reflect
psychologically different processes.
Aptitude X Treatment Interactions
Aptitude X treatment interactions were evaluated
by comparison of regression slopes for different treat-
ments, using F tests for heterogeneity of regression.
Tests were subsequently performed to determine the signif-
icance of individual regression slopes for each treatment.
Inherent in the basic theoretical considerations
underlying the investigation of aptitude X treatment inter-
actions is educational adaptation to individual differences
in the learner. Instruction can be adapted only if there
are alternative treatments leading to the same terminal
objective, and only if the regression of criterion scores
on aptitude scores have the disordinal pattern illustrated
in Figure 6 in which the regression line relating aptitude
93
to criterion scores under one treatment intersects the
regression line for the alternative treatment (Cronbach,
1965). Consequently, while the results of all regression
Criterion
Treatment A
Treatment B
Aptitude
Figure 6. Disordinal Interaction
analyses computed are reported, only the results of analyses
in which significant disordinal interactions occurred have
been presented in either graphic form or subsequent dis-
cussion.
A preliminary step in evaluating aptitude
treatment interactions was to compute regression equations
for all ability measures, using scores obtained for both
written measures. Of 48 F tests computed to test hetero-
geneity of regression, 6 significant interactions were
found. This step was taken in order to improve procedures
for selecting variables for the analysis of aptitude X
treatment interactions for classroom performance measures.
Of 21 tests for heterogeneity of regression subsequently
94
computed for the classroom performance measures, 7 signifi-
cant disordinal interactions were found. Thus a consider-
able improvement in selection can be observed.
However, it should be noted that the latter fig-
ure is somewhat arbitrarily low. Tests for heterogeneity
of regression were computed for T1 as well as for T2 and T3,
even though it was clear that significant interactions
were to be found only in the later stages. Moreover, when
it appeared that the strength of a relationship between a
total test score and a performance measures was largely
accounted for by the relationship of one of the part scores
to that measure, only the part score was used for regres-
sion analysis, even though in some cases the total score
would also have produced a significant disordinal inter-
action.
Classroom Performance Measures
Analytic Questions. Results of regression analy-
ses of aptitude x treatment interactions, using the fre-
quency of Analytic Questions as the criterion measure, are
summarized in Table 25. These analyses show that scores
on Hidden Figures--Part I produced significant disordinal
TABLE 25
SIMPLE REGRESSION ANALYSES FOR ANALYTIC QUESTIONS
PredictionTeachingSession
Treatment Group
FFilm-Mediated
Modelinga b
WrittenModelinga b
Controla b
Hidden Figures - Part 1 T1
3.44 .03 2.05 .24 2.90 .14 .64
Hidden Figures Part 1 T2
14.90 -.24 4.19 .61** 3.34 .02 4.02*
Hidden Figures - Part 1 T3
16.08 -.3.i)* 3.87 .96** 2.71 .05 8.25**
Film Memory T2
10.60 .14 6.12 .10 -.47 .18 .04
Film Memory T3
.96 .62 9.14 .05 -2.57 .28 1.83
Sentence Reproduction-Part 1 T2
2.66 .08 -6.66 .11 6.22 -.02 1.19
Sentence Reproduction -Fart 1 T3
19.40 -.04 -2.50 .09 1.63 .01 1.37
*p<.05
**p.01
Note: a and b are of the form: 7 box
96
2 4 6 8 10 12 14 16
HIDDEN FIGURES -PART I
FIGURE 7. INTERACTION OF HIDDEN FIGURES - ISCORES WITH INSTRUCTIONAL TREAT-MENTS FOR ANALYTIC QUESTIONS.
interactions with the frequency of Anal tic Questions both
for T2 (F = 4.02; p < .05) and for T3 (F = 8.25; p < .01);
while the interaction for T1 was insignificant (F = .64).
In each case, scores for Hidden Figures were
positively related to performance in the Written Modeling
condition while negatively related to performance in the
Film-Mediated Modeling condition. Thus high ability Ss
learned to use Analytic Questions better from the Written
Modeling treatment than from the Film-Mediated Modeling
treatment, while those scoring low on Hidden Figures profited
more from the Film-Mediated treatment. Hidden Figures
scores did not display consistent relationship to per-
formance in the Control group. It should be noted that
the magnitude of the interactions increased across teaching
sessions. The obtained relationship between scores on
Hidden Figures--Part I and the frequency of Analytic Ques-
tions for the three experimental conditions is illustrated
in Figure 7.
Categories of Analytic Questions Aptitude X
treatment interactions obtained using the variety of cate-
gories of Analytic Questions as dependent variable are
shown in Table 26. The results of these regression analyses
TABLE 26
SIMPLE REGRESSION ANALYSES FOR CATEGORIES OF ANALYTIC QUESTIONS
Prediction
TeachingSession
Treatment Group
Film-MediatedModelinga b
WrittenModelinga b
Controla b
Hidden Figures-Part 1 T1
1.29 .02 1.10 .10 1.64 .02
Hidden Figures-Part 1 T2
4.11 -.06 2.14 .12* 1.76 -.04
Hidden Figures-Part 1 T3
4.76 -.14** 2.08 .14* 1.30 .01
Film Memory T2
.09 .07 1.49 .07 .09 .07
Film Memory T3
1.05 .12 2.11 .04 -1.05 .12
Sentence Reproduction-Part 1 T2
3.77 .00 -.51 .02 4.06 -.02
Sentence Reproduction-Part 1 T3
5.91 -.02 1.16 .01 1.22 .00
*p<.05
**p<.01
No :e: a and b are of the form: Y = a + bx
F
.97
3.28*
8.88**
.63
.78
2.37
1.54
99
..................................................................................................................
4 6 8 10 12 14 16
HIDDEN FIGURES - PART I
FIGURE 8. INTERACTION OF HIDDEN FIGURES - ISCORES WITH INSTRUCTIONAL TREAT-MENTS FOR CATEGORIES OF ANALYTICQUESTIONS.
100
disclose the interaction of scores on Hidden Figures- -
Part I with instructional treatment conditions both for
T2 (F = 3.28; p < .05) and for T3 (F = 8.88; p < .01).
Hidden Figures scores did not produce significant inter-
action for T1 (F = .97).
Scores for Hidden Figures were positively related
to performance in the Written Modeling condition, while
in the Film-Mediated Modeling condition, a negative re-
lationship was obtained between aptitude and outcome.
Subjects scoring high on Hidden Figures learned to use a
greater variety of Analytic Questions from the Written
Modeling condition; those scoring low benefited more from
the Film-Mediated Modeling condition. Hidden Figures scores
showed little systematic relation to performance in the
Control group. The magnitude of the interactions again
displayed a tendency to increase across performance trials.
These interaction effects are illustrated in Figure 8.
High Quality Analytic Questions The results
of regression analyses obtained with the quality of Analytic
Questions as the criterion measure are presented in Table
27. These results show that scores on both. Hidden Figures--
Part I and Film Memory produced significant disordinal
interactions with the quality of Analytic Questions used.
1
TABLE 27
SIMPLE REGRESSION ANALYSES FOR HIGH .QUALITY ANALYTIC QUESTIONS
Prediction
Treatment Group
Teaching Film-Mediated WrittenSession Modeling Modeling Control F
a b a b a b
Hidden Figures -Part 1
Hidden Figures-Part 1
Hidden Figures- Part 1
Film Memory
Film Memory
Film Memory
Maze Tracing - Part 2
Maze Tracing-Part 2
Maze Tracing-Part 2
T1
1.83 .01 1.43 .10 1.43 .12 .49
T2
11.54 -.26 2.57 .40* 1.63 .04 4.2 0*
T3
11.89 -.27 3.01 .57** 1.88 -.03 5.70*i
T1
2.02 -.01 1.96 .00 2.52 -.02 .01
T2
8.53 .07 4.49 .03 -1.02 .14 .10
T3
-2.38 .62** 7.16 -.02 -.85 .13 4.42*
T1
.21 .12 2.86 -.06 1.33 .06 1.19
T2
12.64 -.20 2.49 .19 .50 .09 1.76
T3
11.61 -.11 2.41 .31 -.48 .15 1.44
*p<.05
**p.01
Note: a and b are of the form: Pi= a + bx
SNOLLS3110 011A1VNV Ainvno HOIH t:104
-143111 1VN011311N.LSNI H11M S3803S I - s3an9u N3001H 40 N01.1.31/831N1 '6 3811914
I leiVd-S3N11914 N300IH 91 01 31 01 8 9 t7 3
1\1111111
ZOT
103
cf) 20zO
w 16a
i= 12
4z4 8
7.144
cD
E
8 12 16 20 24 28 32FILM MEMORY
FIGURE 10. INTERACTION OF FILM MEMORY SCORESWITH INSTRUCTIONAL TREATMENTSFOR HIGH QUALITY ANALYTICQUESTIONS
104
Hidden Figures scores interacted significantly
with the presentation conditions both for T2 (F = 4.20;
p < .05), and for T3 (F = 5.70; p < .01), while the inter-
action for T1 was insignificant (F = .49). These results
were highly similar to those obtained for the frequency
and variety of Analytic Questions, with scores for Hidden
Figures displaying a positive relationship to performance
in the Written Modeling condition and a negative relation-
ship to performance in the Film-Mediated Modeling condition.
Again, Ss with high scores on Hidden Figures learned to
use high quality Analytic Questions better from the Written
Modeling condition, while those scoring low performed better
in the Film-Mediated Modeling condition. Hidden Figures
scores showed little relation to performance in the Control
group. In addition, the interactions tended to increase
as instruction progressed. The relationships obtained
are depicted in Figure 9.
Scores for Film Memory produced a significant
disordinal interaction only for T3 (F = 4.42; p < .05).
Tests for heterogeneity of regression did not disclose
significant interaction either for T1 (F = .01) or for
T2
(F = .10). For T3 Film Memory scores were positively
related to performance in the Film-Mediated Modeling condi-
tion, while unrelated to performance in either the Written
105
Modeling condition or the Control group. These results
show that Ss with high scores on Film Memory learned to
use high quality Analytic Questions better from the Film-
Mediated Modeling treatment, whereas low scoring Ss profited
more from the Written Modeling Treatment. These interaction
effects are shown in Figure 10. In contrast to the inter-
actions previously discussed, this interaction appeared
on T3 with no observable trend in this direction in the
previous teaching sessions.
Written Measures
True-False Test Analyses of the interaction of
ability measures with treatment conditions for scores on
the True-False Test showed that scores for both Film Memory
and Maze Tracing--Part II produced significant disordinal
interactions (Table 28). Although Film Memory scores in-
teracted significantly (F = 4.01; p < .05) with scores
on the True-False Test, the direction of differences was
in contrast to those reported for Film Memory using the
quality of Analytic Questions as the dependent measure.
This time, the regression slope obtained for the Written
Modeling condition was positive, while a negative relation
TABLE 28
SIMPLE REGRESSION ANALYSES FOR TRUE-FALSE TEST
Treatment Grou
FPredictionFilm-MediatedModelinga b
WrittenModelinga
Control
Hidden Figures - Part I 10.06 .11 10.63 .14 7.03 .01 1.03
Hidden Figures - Part II 9.57 .19 9.88 -.02 7.00 .02 .41
Hidden Figures Total 9.62 .09 10.45 -.05 7.00 .00 1.55Identical Pictures - Part I 8.19 .06 10.56 -.02 8.13 -.03 1.10Identical Pictures - Part II 7.81 .07 9.37 .00 6.79 .00 .71
Identical Pictures - Total 7.70 .04 9.71 .00 7.80 .00 1.02
Maze Tracing - Part I 10.75 .00 8.10 .15 11.77 .12 .99
Maze Tracing - Part II 10.74 .00 5.91 .27** 8.58 -.10 4.94**
Maze Tracing - Total 10.74 .00 6.44 .13* 8.14 -.04 3.11*Word Arrangement - Part I 10.17 .03 9.38 .02 5.70 .07 .16
Word Arrangement - Part II 9.13 .07 9.64 .00 6.04 .05 .27
Word Arrangement - Total 9.46 .03 9.46 .00 5.88 .03 .04
Advanced Vocabulary - Part I 9.48 .09 10.99 -.09 7.65 -.04 .59
Advanced Vocabulary - Part II 7.91 .20 9.81 -.00 6.39 .05 .85
Advanced Vocabulary - Total 8.34 .09 10.57 -.03 6.83 .00 .85
Verbal GRE 7.29 .00 10.77 .00 6.48 .00 1.51Film Memory 13.09 -.11 6.32 .17* 3.65 .17* 4.01*
Memory for Ideas - Part I 9.15 .07 11.49 -.08 6.62 .02 2.53
Memory for Ideas - Part II 10.40 .02 9.28 .03 6.78 .02 .00
Memory for Ideas - Total 9.38 .04 10.17 -.01 6.53 .02 .53
Sentence Reproduction-Part I 11.67 -.00 4.56 .04 3.73 .02 1.25
Sentence Reproduction-Part II 8.51 .02 8.81 .00 7.83 .00 .47
Sentence Reproduction-Total 9.20 .00 6.31 .01 6.23 .00 .19
*p<.05**p<.01
Note: a and b are of the form: Ai = a + bx
ILI
Film - MediatedModeling
WrittenModeling
Control
10 15 20 25FILM MEMORY
30 35
FIGURE II. INTERACTION OF FILM MEMORYSCORES WITH INSTRUCTIONALTREATMENTS FOR TRUE - FALSETEST SCORES
14
12
WrittenModeling
Film - Mediated1.. 10 Modelingcowi-w 8Cl)J4U.
1 6Wocc
4
Control
4 8 12 14 20MAZE TRACING - PART II
108
24
FIGURE 12. INTERACTION OF MAZE TRACING -a SCORES WITH INSTRUCTIONALTREATMENTS FOR TRUE -FALSETEST SCORES
109
was obtained between aptitude and outcome in the Film-
Mediated Modeling condition. Thus, high ability Ss learned
to identify categories of Analytic Questions better from
the Written Modeling condition; those scoring low on Film
Memory performed better in the Film-Mediated Modeling con-
dition.
Sao es for Maze Tracing were positively related
to performance in the Written Modeling condition, while
displaying no :elationship to performance in the Film-
Mediated Modeling condition. Subjects scoring high on
Maze Tracing learned to identify categories of Analytic
Questions better from the Written Modeling condition, while
low scoring Ss benefited more from the Film-Mediated Model-
ing condition. The interaction effects obtained are de-
picted in Figures 11 and 12.
Matching Test The evaluation of aptitude
treatment interactions for scores on the Matching Test
showed that both Part II and Total scores for Maze Tracing
displayed significant disordinal interactions (Table 29).
While both scores for Maze Tracing produced significant
interactions, it is clear that the interaction for the
Total score (F = 5.64; p < .01) is largely accounted for
TABLE 29
SIMPLE REGRESSION2tANALYSES FOR MATCHING TEST
Predictcx
Treatment Group
FFilm-Mediated
Modeling
ab
Written
Modeling
ab
Control
a
Hidden Figures - Part I
11.50
.07
8.84
.26*
8.78
-.12
3.33*
Hidden Figures - Part II
11.25
.11
10.49
.00
8.62
-.09
.50
Hidden Figures - Total
11 26
.05
9.30
.09
8.81
-.06
1.72
Identical Pictures - Part I
11.52
.00
3.47
.17
5.51
.06
1.38
Identical Pictures - Part II
11.15
.02
4.95
.13
9.08
-.03
1.53
Identical Pictures - Total
11.24
.00
3.89
.08
7.26
.00
1.27
Maze Tracing - Part I
13.44
-.15
7.69
.25
9.21
-.11
2.46
Yaze Tracing - Part II
12.95
-.08
3.77
.47**
9.09
-.07
7.44**
Maze Tracing - Total
13.40
-.06
4.74
.22**
9.29
-.05
5.64**
Word Arrangement - Part I
10.91
.05
9.87
.03
7.43
.03
.12
Word Arrangement - Part II
9.96
.09
10.70
.00
8.11
.00
1.26
Word Arrangment - Total
10.19
.04
10.51
.00
7.77
.00
.78
Advanced Vocabulary - Part I
9.20
.20
9.23
.09
12.46
-.32
2.43
Advanced Vocabulary - Part II
8.44
.24
7.56
.21
8.76
-.05
1.20
Advanced Vocabulary - Total
8.28
.13
8.44
.07
10.51
-.09
1.83
Verbal GEE
5.52
.01
9.31
.00
7.79
.00
1.08
Film Memory
11.37
.03
7.44
.15
4.67
.17
.36
Memory for Ideas - Part I
11.36
.02
9.69
.04
6.01
.09
.26
Memory for Ideas - Part II
11.17
.05
9.34
.07
6.83
.09
.03
Memory for Ideas - Total
11.03
.02
8.11
.06
5.74
.06
.28
Sentence Reproduction - Part
I6.63
.04
2.99
.05
9.11
.00
.93
Sentence Reproduction - Part
II
9.57
.02
7.03
.02
4.87
.02
.05
Sentence Reproduction - Total
7.91
.01
4.29
.02
5.47
.00
.19
*p<.05
**p<.01
Note:
a and b are of the form:
7= a + bx
H 0
111
WrittenModeling
Film MediatedModeling
I0
co
I.- 8CD
6
2
Control
4 8 12 16 20MAZE TRACING - PART /I
24
FIGURE 13. INTERACTION OF MAZE TRACING -ItSCORES WITH INSTRUCTIONALTREATMENTS FOR MATCHING TESTSCORES
112
by the strength of the interaction for Part II scores
(F = 7.44; p < .01). Consequently, since the results are
substantially the same for both scores, only the inter-
action for Part II scores are presented for further dis-
cussion and graphic illustration.
Similar to the results for the True-False Test,
a positive relationship was obtained between aptitude and
criterion in the Written Modeling condition, while scores
for Maze Tracing were unrelated to performance in the Film-
Mediated Modeling condition and the Control group. Again,
Ss with high scores on Maze Tracing learned to identify
specific instances of Analytic Questions better from the
Written Modeling treatment, whereas those scoring low prof-
ited more from the Film-Mediated Modeling treatment. This
relationship is illustrated in Figure 13.
CHAPTER IV
DISCUSSION OF RESULTS AND
IMPLICATIONS FOR FUTURE RESEARCH
Summary of Data: Hypotheses Tests
This study sought to examine the effects of verbal
and perceptual dimensions of individual differences in re-
lation to the efficacy of two different kinds of modeling
proc(Idures in the acquisition of a teaching skill. The
basic hypotheses tested were:
1. Both Film-Mediated and Written Modeling conditions
will produce significantly greater changes in the
response strength of desired behaviors than will
the Control condition.
2. For subjects receiving the Film-Mediated Modeling
condition, criterion scores should show stronger
relation to perceptual abilities than for subjects
receiving the Written Modeling condition.
3. For subjects receiving the Written Modeling con-
dition, criterion scores should show stronger re-
lation to verbal abilities than for subjects
receiving the Film-Mediated Modeling condition.
113
114
The general premises from which these hypotheses
were derived were:
1. The rate and level of learning of a given teaching
strategy varies as a function of model presenta-
tion.
2. The effectiveness of instructionsamethod.s varies
from subject to subject, with such differences
being correlated with trainee aptitudes.
Specific predictions, though tentative, were based
on theoretical considerations which suggested that the re-
quirements of the Written and Film-Mediated Modeling con-
ditions were sufficiently different to produce different
ability-performance relationships.
Instructional Treatment Main Effects
Support for Hypothesis 1 is dependent upon sig-
nificant between-group differences in measures of the de-
pendent variable. The appropriate statistical tests of
this hypothesis are therefore the comparisons of pairs of
treatments following an overall significant F ratio.
Data in Tables 9 through 18 strongly support
this hypothesis. Following initial written instructions
115
common to all groups, both Written and Film-Mediated Model-
ing conditions led to significantly higher frequency, va-
riety and quality of Analytic Questioning than did the
Control group, and thus to a significantly lower frequency
of Non analytic Questions also. Similarly, wiTh respect
to the written measures, Ss in the Written and Film-Mediated
Modeling conditions matched and identified significantly
more correct items than did Control group Ss. Significant
differences existed between the two modeling conditions and
the Control group for all measures of the dependent variable.
It should again be noted that while statistical
consideration has been given to several measures of the de-
pendent variable, the strength of association among these
measures (Table 7 ) in some cases is sufficiently high to
suggest that these measures do not represent psychologically
different variables. However, these variables were ana-
lyzed separately in view of differences which existed in
ability-performance relationships on some occasions, and
because of the unavailability of multivariate analysis of
variance techniques.
Additional support for Hypothesis 1 is provided
by within group analyses of the changes in Analytic Ques-
tioning behavior from base rate to subsequent teaching
sessions. This information is provided in Tables 9 through
18. Both Film-Mediated and Written Modeling conditions
produced significant increases in frequency, variety and
quality of Analytic Questioning., while the Control group
did not display such increases. Moreover, both Written and
Film-Mediated Modeling conditions produced significant de-
creases in the frequency cf Nonanal tic Questioning, while
the Control group did not display such decrements.
The nature of the dependent variable appears
to have been an important factor in the results obtained.
It has been observed that modeling procedures are most
effective in transmitting response patterns which are
new or weakly established (Bandura 1965, 1966). Frequently,
viable teaching skills and strategies, because of their
demonstrable effectiveness in the classroom, can be ex-
pected to be occurring in some strength prior to treat-
ment. This raises the question of the utility of modeling
procedures for training purposes in such instances. How-
ever, the occurrence of Analytic Questioning prior to treat-
ment was found to be surprisingly infrequent. Even follow-
ing exposure to set induction materials describing the na-
ture af analysis, subsequent questioning was most frequently
evaluative rather than analytic. Consequently, the low initial'
117
operant strength of Analytic questioning appears to have
provided optimal conditions 2or demonstrating the effective-
ness of the modeling procedures.
A most striking finding was the consistent su-
periority of the Film-Mediated Modeling condition across
all measures of the dependent variable. No specific pre-
dictions had been formulated concerning differences exist-
ing between 'he Written and Film-Mediated Modeling condi-
tions. While both of these methods have been effective as
training procedures, differences between the two treatments
have not consistently been found (Bandura and Mischel,
1965, McDonald and Allen, 1967, McDonald, 1968). However,
systematic feedback and reinforcement have often been
included in these studies. While the inclusion of these'
variables has generally contributed to strengthening both
modeling treatments, it has perhaps obscured the specific
contributions of the observational treatment variable.
This interpretation of the lack of consistency in
these results would appear to conform to Bandura's argument
that contiguity accounts for the acquisition of matching
responses, whereas reinforcement influences the performance
of imitatively learned responses (Bandura, 1965). Some
evidence on this point has been provided in an experiment
118
in which the introduction of positive incentives com-
pletely wiped out previously observed performance differ-
ences (Bandura, 1965). In contrast, the exclusion of
feedback and reinforcement variables in the present study
permitted investigation of the specific effects of an ob-
servational experience on the acquisition of a teaching
skill.
The clear cut differences between the Written
and Film-Mediated Modeling conditions across all measures
of the dependent variable supports Carroll's (1959) conten-
tion that:
Psychologists have too often confused the spoken andthe written word, or at least they have assumed toofreely that spoken and written words are equivalentstimuli (p. 110).
A typed transcript differs from a filmed portrayal in sev-
eral important respects. A detailed discussion of these
differences was given in Chapter I. In addition to pro-
viding a behavioral conception of both the teaching skill
and lesson pace, the film-mediated portrayal standardizes
intonation and stress patterns which subjects may implicitly
impose. Differences of this nature may prove to be of
considerable importance. Thus, an actual performance is
119
likely to provide substantially more relevant cues with
greater clarity than can be conveyed by a verbal descrip-
tion. From the average data alone then, training under
Film-Mediated Modeling conditions appears to have been
generally more efficient.
Aptitude X Treatment Interactions
Hypotheses 2 and 3 imply disordinal interaction
between aptitude and treatment--more specifically, that
the regression line relating aptitude to criterion scores
under one treatment intersects the regression line for the
alternative treatment. The statistical tests of these hy-
potheses are F tests for heterogeneity of regression.
The evidence with respect to Hypotheses 2 and
3 is mixed. Although aptitude x treatment interactions
were obtained, the direction of the interactions did not
consistently correspond to predictions. Analyses of apti-
tude X treatment interactions indicated that scores on Film
Memory interacted significantly with the quality of Ana-
lytic Questions. Scores for Film Memory were positively
related to performance in the Film-Mediated Modeling con-
dition, while unrelated to performance in the Written
120
Modeling condition. Subjects scoring high on the Film
Memory Test learned to use high quality Analytic Questions
better from the Film-Mediated Modeling treatment, while
those scoring low profited more from the Written Modeling
treatment. These results are consistent with Hypothesis 2.
In contrast to these findings however, scores
on Hidden Figures produced significant disordinal inter-
actions with the frequency, variety and quality of Analytic
Questions. While scores on Hidden Figures were positively
related to performance in the Written Modeling condition,
they were negatively related to performance in the Film-
Mediated Modeling condition. Thus high ability Ss learned
to use greater frequency, variety and quality of Analytic
Questions from the Written Modeling condition than from
the Film-Mediated Modeling condition; those scoring low
on Hidden Figures learned better from the Film-Mediated
Modeling condition. Similarly, with respect to performance
on the written measures, scores on Maze Tracing and Film
Memory interacted with presentation conditions. Using
Film Memory scores, the regression slope obtained for
the Written Modeling condition was positive while a nega-
tive relation was obtained between aptitude and outcome
in the Film-Mediated Modeling condition. Scores for Maze
121
Tracing were positively related to performance in the Writ-
ten Modeling condition, while unrelated to performance in
the Film-Mediated Modeling condition. These results show
that Ss with high scores on Film Memory and Maze Tracing
performed better on the written measures in the Written
Modeling treatment, whereas low ability Ss benefited more
from the Film'-Mediated Modeling treatment. These results
are in sharp contrast to those predicted. Consequently,
Hypotheses 2 and 3 are unsupported by these data. While
aptitude x treatment interactions were obtained, the direc-
tion of the interactions was not consistent and at times
opposite to that hypothesized.
Whenever differences result from a study in con-
tradiction to the hypotheses stated, it is incumbent upon
the experimenter to offer post hoc explanation for such dif-
ferences, although it is not possible, of course, to verify
such speculatory explanation within the scope of this study.
The initial hypotheses were based upon an analysis
of learning tasks and processes corresponding to a model
proposed by Melton (1967) for the investigation of learning
and individual differences. An inspection of the indi-
vidual differences believed to enter at each point in the
learning process can be seen in Figure 14. The organization
rm (sm )
Si1 1(s ) >RaR
Re
* Hidden Figures* Maze Tracing
Identical PicturesVerbal ComprehensionVerbal GRE
122
* Film MemoryMemory for IdeasSentence ReproductionExpressional Fluency
Figure 14. Organization of AbilityVariables in a Multi-Process Model of Learning
of ability measures within the model has been somewhat
arbitrary. Measures such as the GRE Verbal Aptitude scores,
with both comprehension and reasoning components, are
sufficiently diverse to suggest that there is reason to
consider alternative placement. Other measures may also
be characterized by similar overlap. However, it should
be emphasized that this model has been used as a heuristic
device in attempting to select and organize task and ability
variables. It is not intended that the present organization
be accepted as definitive. Further research will be re-
quired to clarify the use of this model.
Of nine ability variables studied, three were
found to interact with presentation conditions. If the
direction of the interactions is examined with respect to
the point in the learning process at which they are believed
t23
to enter, certain trends are suggested. Accordingly, it
is suggested that such trends may conceivably be interpreted
within the framework of the model.
It is becoming increasingly apparent that hypoth-
eses concerning the direction of aptitude g treatment
interactions cannot be based exclusively upon superficial
content similarities between aptitude tests and learning
tasks. Visual or verbal modes of instructional presen-
tation may or may not be related to corresponding scores
on perceptual or verbal aptitude tests.
While it would appear thAt a perceptual mode of
presentation would constitute a demand on perceptual en-
coding systems, the audio-visual presentation could also
conceivably serve a compensatory function through the pro-
vision of perceptual information that might otherwise be
demanded of the subject. Similarly, the lack of audio-
visual content in instructional presentation may require
subjects to generate their own perceptual detail, thus
constituting a demand on the perceptual encoding system.
The stimulus differentiation component of the
model is specifically concerned with the encoding process.
Of the five ability variables studied here, two were found
to interact with presentation conditions. An inspection of
124
the direction of these interactions can be interpreted to
suggest that those abilities involved in the stimulus dif-
ferentiation component may generally serve the compensa-
tory function described above. The purely verbal presen-
tation of a behavioral sequence may require subjects to
generate a behavioral conception of that sequence, thus
creating a perceptual demand. Conversely, an audio-visual
presentation, rich in perceptual detail may require an ab-
straction of a verbal conception, constituting a verbal
requirement.
This compensatory function may be illustrated
in the case of those interactions involving Hidden Figures.
Performance in this test appears to require subjects to
visualize the configurations they are looking for within
the context of a given perceptual field. Knowledge of
the exact configurations which have to be kept in mind
is specially emphasized. Within the Film-Mediated Modeling
condition perceptual schema are provided. Ss low in the
ability to generate such information for themselves are
compensated by this treatment and thereby improve their
performance. However, for Ss who are able to generate such
schema for themselves, the presentation of such information
at a predetermined pace may not only fail to be facilitative.
125
but may inhibit or interfere with the generation of per-
ceptual information or produce frustration and boredom.
An interesting observation was that learners with high
scores on Hidden Figures were actually hindered in the
Film-Mediated Modeling condition, as demonstrated by the
negative relationship between aptitude and performance.
The presence of such negative ability-performance relation-
ships may indicate that reliance on a strong ability may
lead a learner to choose a strategy which unwisely attempts
to exploit that ability in a situation where information
is available which would make a different strategy more
effective (Bunderson, 1968).
In the Written Modeling condition, where per-
ceptual schema are not provided, Ss high in the ability
to generate such information appear to do well, possibly
because following their own pace, they can freely generate
the needed perceptual information. In contrast, Ss low
in ability to generate perceptual information for them-
selves are not provided with this information in the Written
Modeling condition, and subsequently perform less well.
Because of the ',connection of Hidden Figures
with both generalized intellectual ability (Botzum, 1951;
Pemberton, 1952) and with a series of tests of a trait
126
called "field independence" (Witkin et al., 1952), there
is reason to consider possible alternative interpretations
of these results. However, the alternative interpreta-
tions are not necessarily inconsistent with the present
interpretation. Witkin et al. (1962), in a review of the
relevant literature concluded:
These studies provide impressive support for theview that flexibility of closure and field-independencemay be different names for the same dimension. (p. 52)
These findings support Thurstone's (1944) contention that
flexibility of closure may be found to represent parameters
which transcend the immediate perceptual content in terms
of which it had tentatively been identified.
Interactions involving Maze Tracing scores are
believed to be more closely related to the pacing factor.
Maze Tracing requires speed in scanning a complicated
spatial field. In this case, the pace of the film-mediated
model may have served as an equalizer in which the pre-
determined pace of the model imposed a ceiling on the
rate at which Ss high in this ability could scan and process
information, while being sufficiently slow to allow those
Ss low in this ability to profit thereby. Consequently,
the correlation between Maze Tracing and performance
in the Film-Mediated Modeling condition was atten-
uated.
127
In addition, the written model can be charac-
terized as a spatial configuration as well as a verbal
configuration. Teacher discourse was identified and sepa-
rated from student discourse. Questions were set off
by question marks and were generally located in certain
positions within a paragraph of teacher discourse. Cue
discrimination was provided in upper case letters marked
by parentheses. Ss in the Written Modeling condition
could determine their own pace without limitations im-
posed on the speed at which this array of information
could be scanned and reviewed. Accordingly, Ss high in
spatial scanning ability could quickly review this writ-
ten material and locate points at which critical infor-
mation (i.e., cue discrimination and specific examples
of Analytic Questions) was given and thereby improve their
level of performance. Low ability Ss might be less able
to take advantage of these features.
While these two ability variables are the only
variables in the stimulus differentiation component of the
model which interacted significantly with instructional
presentation conditions, patterns of correlations with re-
spect to other ability variables in the stimulus differ-
entiation component are not inconsistent with this
interpretation. Although verbal abilities did not inter-
act significantly with the presentation conditions, the
possibility exists that the cue discrimination, in which
appropriate verbal labels were provided for critical be-
havioral sequences served to attenuate the relationship
between verbal abilities and performance.
Findings such as these on the relationship of
test to task are not unique in the literature on aptitude
X treatment interactions (Blaine and Dunham, 1968). Re-
sults suggesting a compensatory function of certain task
elements have occurred sufficiently often to warrant con-
sidering these data relevant to speculations on the rela-
tionship of a given aptitude to a particular phase in a
learning task.
In contrast to these findings, are the results
obtained regarding the response integration component of
the model. Of the four abilities studied here, two apti-
tude X treatment interactions were disclosed. Both aptitude
x treatment interactions involved Film Memory in which the
129
findings were in direct contradiction to one another. It
will be recalled that the quality of Analytic Questions used
was positively related to FilmIMemory in the Film-Mediated
Modeling condition and unrelated to performance in the
Written Modeling Condition. Conversely, using scores on
the T-F test as the criterion measure, Film Memory was
positively related to performance in the Written Modeling
condition and negatively related to performance in the Film-
Mediated Modeling condition. Although the criteria are
clearly different, such apparently contradictory results
provide an unsatisfactory basis for conclusions regarding
the relationship between the response integration component
of the model and the expected direction of aptitude x treat-
ment interactions. However, a supplemental study of another
set of dependent measures was undertaken. While these
measures were not a part of the main investigation, and thus
will not be reported in detail, the results obtained are rele-
vant to clarification of ambiguities in the data reported.
The supplemental set of dependent measures con-
sisted of a recall test in which Ss were asked to list in
writing as many Analytic Questions as possible, based on
their lesson material, within a 10 minute period. This
was collected along with the other two written measures
44
130
reported) and was rated in a manner identical to that of
the classroom performance measures. Analyses of aptitude
X treatment interactions for these data disclosed that
scores on Film Memory interacted significantly with the
frequency and quality of Analytic Questions. In both cases,
scores on Film Memory were positively related to performance
in the Film-Mediated Modeling condition and unrelated to
performance in the Written Modeling condition. Thus Ss
scoring high on Film Memory asked higher frequency and
quality of Analytic Questions in the Film-Mediated Modeling
treatment) while low scoring Ss performed better in the
Written Modeling treatment. Similarly) scores on Sentence
Reproduction-Part II interacted significantly with the
variety of categories of Analytic Questions used. While
the regression slope obtained for the Film-Mediated Modeling
condition was positive) scores on Sentence Reproduction were
unrelated to performance in the Written Modeling condition.
Subjects with high scores on Sentence Reproduction learned
to ask a greater variety of Analytic Questions from the
Film-Mediated modeling condition; those scoring low bene-
fited more from the Written Modeling condition. These
findings suggest that in contrast to aptitude X treatment
interactions involved in the stimulus differentiation
131
component, aptitude X treatment interactions involving
the response integration and memory components of the
model may generally be consistent with earlier predictions- -
more specifically, the relationship of test to task appears
direct rather than compensatory, with scores on tests of
audio-visual memory related to performance in an audio
visual mode of presentation, while unrelated to performance
in a written mode of presentation.
It is recognized that this interpretation cannot
accommodate the relationship between Film Memory scores
and performance on the T-F test. At this time the experi-
menter is unable to provide a satisfactory explanation
for these results on the basis of present data. While
it is possible that this finding may be due to chance, an
alternative explanation might center on the nature of the
task itself. The discriminations required in the T-F
test are much coarser than those required for other de-
pendent measures. Conversely, the discriminations required
on the Film Memory test are quite fine. An examination
of the pattern of correlations suggests that the relation-
ship of ability measures to performance on the T-F test
is somewhat different from those obtained with other mea-
sures of the dependent variable. It has been suggested
(Fitts, 1965) that the perception and encoding of stimuli
132
depend in part upon the required response mode. However,
the precise effects of such differences in the nature of
a task remain a suitable topic for future investigation.
Information concerning the associational-
mediational component of the model is limited. Within
the context of this study there were no available measures
of ability to generate associative context, analogies,
or verbal-behavioral transformation. However, an experi-
mental measure was developed for use in collecting sup-
plemental data which attempted to assess the ability to
abstract a verbal conception from a behavioral representa-
tion. While scores on this test did not interact signif-
icantly with performance on any of the measures reported in
the main investigation, these scores did interact signif-
icantly with the frequency of Analytic Questions on the
recall measure, disclosing a positive relationship to
performance in the Film-Mediated Modeling condition and
a negative relationship to performance in the Written
Modeling condition. These results correspond to those
obtained for the response integration component of the
model. It should be strongly emphasized, however, that
the limited tryout of both content and procedures in ad-
dition to the lack of reliability and validity data for
this test prohibit the use of this information as anything
other than heuristic.
133
In summary, it has been suggested that the model
proposed by Melton (1967) for the investigation of learn-
ing and individual differences might be used as a plausible
basis for predicting the relationships among learner charac-
teristics and learning outcomes under different instruc-
tional procedures, in the manner described above. While
the evidence on this matter is not overwhelming, and may
be variously interpreted, the model would appear to be
useful, at least in a heuristic sense.
An additional finding of considerable interest
was the tendency for aptitude X treatment interactions
to increase in magnitude across performance trials. This
occurred for all interactions involving classroom per-
formance measures. These findings are in contrast to
those reported by Woodrow (1938), Fleishman and his as-
sociates (1965) and others (Gagne and Paradise, 1961) in
which the contribution of cognitive factors in a learning
task tended to decrease with practice. This has generally
been true both for intellectual and psychomotor tasks.
There are several reasons for such differences.
First of all, the type of task that has generally been
used has been of relatively short and simple content.
This may serve to impose a ceiling on performance which
134
may be reflected in the decreasing contribution of cogni-
tive factors across learning trials. Conversely, the
present study has investigated ability-performance re-
lationships across trials on an intellectually complex
task. In addition, the conditions of performance in the
present study impose a ceiling only in terms of time rather
than a fixed format with a limited number of problems,
frames, or tasks.
Moreover, this line of investigation has general-
ally been concerned 1.7.1.01 task variables rather than in-
structional variables. Accordingly, ability-performance
relationships have been assessed under conditions of prac-
tice rather than instruction. In contrast, the present
study has investigated ability-performance relationships
under different instructional conditions.
Thus, the combination of the type of task and
instructional conditions may have served to increase the
magnitude of aptitude-treatment interactions across trials.
It should be noted, of course, that the number of trials
was limited. Continuation through further teaching sessions
may eventually have resulted in decreasing ability-
performance relationships.
135
Implications for Future Research
This study represents an initial attempt to
assess the effects of trainee aptitudes on observational
learning in the acquisition of a teaching strategy. The
results obtained suggest that additional research evaluating
differentiation of instruction in teacher training may
be profitable. The surface has barely been scratched in
this area. A variety of relevant research questions are
suggested by the results of this investigation.
With respect to the use of modeling proceaures
in teacher training, several important issues have yet
to be systematically explored. As Cronbach (1965) has
suggested, any meaningful comparison of treatments re-
quires that each treatment be refined into a good repre-
sentative of its kind. Accordingly, we need to look more
closely at the precise nature of effective modeling con-
ditions in order to determine optimal sequencing of in-
struction and number of cues and exemplars provided, as
well as at the effects of varying amounts and types of
set induction and practice. Moreover, the question of
retention and transfer effects of this training has not
yet been examined. A treatment has multiple effects, and
the effectiveness of instructional methods may differ
136
for immediate mastery, retention and transfer. Although
positive preliminary information has been provided on
this point (Bandura and Mischel, 1965) additional research
is required.
There appears to be an abundance of teacher
behaviors which could ostensibly be used as dependent
variables in studies of this type. Anything that a teacher
does or says in an effort to promote student learning
is a potential skill to be learned by prospective teachers.
While the modeling procedures utilized have been successful
in transmitting specific skills such as Analytic Question-
ing, the general validity of these skills for producing
changes in pupil behavior has yet to be demonstrated.
Although many of these skills possess face validity, at-
tention should be given to the relevance of such teacher
variables in terms of their effect on student learning.
With respect to the effects of trainee aptitude
on observational learning, several research questions are
suggested. The results reported in this investigation
indicate that differences in instructional methods facili-
tated learning in some subjects while inhibiting learning
of others, with such differences being related to trainee
aptitudes. It has been suggested that the model proposed
137
by Melton (1967) for investigating individual differences
in learning might be used as a plausible basis for the
prediction of ability-performance relationships. Con-
siderable emphasis has been placed upon the simultaneous
consideration of task and process variables. Clearly,
additional research is required to evaluate the utility
of this model. The extent to which response integration,
stimulus differentiation, coding and a variety of media-
tional processes may be involved in new learning can be
experimentally manipulated. In view of the limited in-
formation obtained in this study with respect to individual
differences in mediational processes, particular attention
should be given in the future to this component of the
model.
Related to the major question of the utility
of the model are several questions concerning the nature
of the specific task utilized in this investigation. While
the task was relatively complex, the difficulty level was
to some extent attentuated through the provision of set
induction and cue discrimintation. Although these features
are believed to have generally strengthened both modeling
conditions under consideration, the question of the extent
to which the provision of this additional information
138
altered ability-performance relationships that would other-
wise be a function of the exclusive effects of an observa-
tional sequence remains unanswered.
In addition, the results of this study have
raised the question of the effect of response mode on
ability-performance relationships. These relationships
differed substantially from classroom performance to writ-
ten measures. It has been suggested (Fitts, 1967) that
the perception and encoding of stimuli depends at least
partially on the responses to be made to the stimulus
information. Accordingly, it is recommended that further
experimental tests of the model include variations in the
response mode used by subjects.
Moreover, the question of ability-performance
relationships involved in retention and transfer measures
on this particular task has yet to be explored. A legiti-
mate question is the extent to which the battery of tests
used samples abilities related to retention and transfer
as well as immediate mastery.
Finally, further investigation of changes in
the contribution of cognitive abilities over practice
would appear to be a potentially fruitful line of investi-
gation. Results such as those obtained in this study
139
indicate that this question cannot be answered simply.
The effects of variables such as the nature of the task,
conditions of instruction or practice, length of the train-
ing period and composition of the subject group appear
worthy of future investigation.
Implications for Educational Practice
The results of this investigation support the
initial premises that the rate and level of learning of
a specific teaching strategy varies as a function of model
presentation; and that the effectiveness of instructional
methods varies from subject to subject, with such dif-
ferences being related to trainee aptitudes.
Although it has not been demonstrated that the
training procedures utilized in this investigation repre-
sent the most effective way to train teachers, these find-
ings provide evidence that through observation, trainees
can acquire principles exemplified in a model's behavior
and use them for generating novel combinations of teaching
behavior. The results of these training procedures were
sufficiently strong and consistent to suggest that they
can be used with reasonable confidence in their effec-
tiveness. Accordingly, this research supports the general
140
recommendation that the use of written and film-mediated
modeling procedures is a highly effective means of modifying
teaching behavior in training contexts analagous to those
described in this experiment. Moreover, while questions
concerning interactions between specific teaching behaviors
and instructional conditions have yet to be resolved,
there is further evidence, from the average data alone,
to suggest that film-mediated modeling procedures possess
cueing properties which tend to recommend their use over
written modeling procedures, at least for establishing
Analytic Questioning. behavior.
This research does not lead to detailed sugges-
tions for specific modification and developments regarding
the individualization of teacher training programs. Ex-
periments such as these have only begun to explore the
wide range of problems concerned with finding effective
teaching techniques for students with different character-
istics. Much research remains to be done before such
recommendations can be made with confidence. However,
there are both cost and efficiency implications for teacher
training in these results. Given replication of these
findings, assignment of trainees to alternating treatment
is appropriate for maximizing learning. Moreover, the
141
cost of training is reduced considerably as more teachers
can be-assigned to written rather than film-mediated treat-
ments.
The results obtained in this study appear suf-
ficiently encouraging to suggest the potential value of
further research evaluating differentiation of instruction
in teacher training. Results such as these, if replicated,
may eventually provide a basis for the individualization
of teacher training programs.
Limitations and Further Statistical Considerations
While the initial purposes of this investigation
have been satisfied, the results obtained generate questions
concerning the most appropriate kinds of statistical anal-
yses for research of this type. The analyses used have
not been exhaustive. More complete statistical elabora-
tion of these results can and should be pursued. A next
logical step would be to combine aptitude variables to
describe complex interactions, using multiple rather than
simple regression analyses.
Additional methodological issues need also to
be considered. While significant disordinal aptitude x
treatment interactions have been obtained, the points
along the ability continuum at which the effects of in-
structional procedures clearly become significantly dif-
ferent have not yet been established. One possibility
for such analysis lies in the application of the Johnson-
Neyman (1936) method in which the continuum of aptitude
test scores can be divided into regions where each in-
structional method is superior.
Moreover, optimal procedures for dealing with
multiple instructional outcomes have not yet been determined.
Simultaneous statistical consideration of multiple out-
comes of instruction is infrequently represented in the
literature. In view of the differential importance of
multiple criteria, it is not yet apparent what weighting
schemes might be most useful when combination is desirable.
It is anticipated that additional analysis of
the present data will continue in an attempt to provide
information useful in resolving these issues.
APPENDICES
143
APPENDIX A
RATER MANUAL
144
145
RATER MANUAL
I. GENERAL RULES
A. When a teacher is eliciting a student response, he is asking aquestion.
B. Rate each inquiry as given. Don't add punctuation or extra wordseven if this seems necessary to "make sense" out of the inquiry.
If inquiries have blanks or data indicating something has beenomitted, rate anyway as long as the inquiry makes sense.
D. Inquiries such as "What else?" "Can you give another example?""Anything else?" "Keep going," "What is he saying?" "What is hegetting at?" "What about?" "How about?" "Any other?" "Is that whathe's saying?" etc., when following an Analytic Question, are con-sidered an extension of that question and are rated as such.
1. An extension takes precedence over any changes of terms.e.g. "What is meant here by garnering glory?" is a definitionalquestion. "Any other reasons?" is considered an extension ofthe definition despite the change in terms.
2. An extension does not refer to a paraphrase, but to the questionpreceding the paraphrase.
E. In case of a conflict of terms use the first term, e.g. "Whatassumptions are made about the reason why people riot?" would berated as an assumption since the assumption precedes the hypothesishere.
F. If a ratable Analytic Question which does not refer specifically tothe written material is embedded in a section of teacher discoursewhich is obviously referring to the article, rate it as Analytic.
G. Questions to be rated as Analytic should be related to the writtenmaterial rather than simply representing the student's opinion.
146
e.g. "What does the author mean by patriotism?" would be rated asAnalytic, whereas "What is your on definition of patriotism?"would not be.
H. Questioning should be directed to critical analysis of the writtenmaterial. This should be interpreted fairly liberally, with thebenefit of the doubt given in borderline cases. The main distinc-tion is that the teacher should be getting at the analysis of thematerial--what has actually been set forth--rather than simplyasking students for their on opinions, value judgments, etc.
I. Be sure to check to see if the teacher discourse at the 12,2 of a
page may be a continuation from the bottom of the preceding page- -
this can make a big difference as to whether the question is to berated as an Analytic Question or not.
J. Be sure to check all of your totals to see that ALL questions havebeen rated. This is VERY important.
II. ANALYTIC QUESTIONS
A. Hypotheses
A hypothesis states a relation of dependency among variables.It predicts that s change in one variable will be accompanied by or willproduce changes in another variable; thus the relationship may be statedas cause and effect or as accompaniment.
Here are some examples of hypotheses: The characteristics ofa college a student attends will be related to the scores he obtains onthe GRE. The first variable is "college characteristics"; the second"examination scores." Attitude toward school will be related positivelyto the amount of biology retained. Attitude toward school is onevariable (since not all students have the same attitude), amount ofbiology retained is another variable. The hypothesis predicts thatpositive attitudes will be associated with greater retention.
Hypotheses are frequently (but not always) stated in the form:If A, then B: If positive attitude toward school, then greater reten-tion of subject matter.
pP
tryry
147
Includes: Any interrogative sentences or statements that ask what
relationship exists between two or more variables.
Examples
1. What is the author's hypothesis?2. What is the problem that is being investigated?3. What does he think is the cause of B?
4. What does he think A causes?5. What is thought here to be the relationship between A and B?
6. How are A and B connected?7. Is X the hypothesis?8. What analogy is he stating?9. Why does he think A happens?
10. What arguments (explanation) does he give for B?
11. How does he account for B?12. What is he trying to prove?
BOTH SIDES OF THE HYPOTHESIS HAVE TO BE RELATED TO THE ARTICLE.
Does not include: What is theme, or main point.
B. Definitions
The word definition comes from the Latin de which means,
"concerning" and "finis," boundary or limits. A definition is therefore
literally a statement concerning the limits or boundaries of the meaning
of a word. Definitions thus have two main tasks to perform: to convey
the essential meaning which is to be the common ground of understanding,
and to mark out its limits with sufficient precision for the purposes
in view.
Includes: Questions requiring students to distinguish the methods and
standards used for defining a word in terms of the written material
they are analyzing (as distinguished from their own opinion or common
sense notion), similarities, differences, definitions by example.
Examples
1. What is meant here by poverty?
2. A proposed civil rights bill would apply to business engaging
in a "substantial" amount of interstate commerce. How is
substantial defined here?
148
3. What is meant here by patriotism?4. How do we know what a hippie is from reading this?5. How does he define objective?6. Does he define :hippie?7. What distinctions does he make between militaristic and
pacifistic patriotism?8. What does he say is the difference between prejudice and
discrimination?9. Who are the hippies? (who here referring to a collective)
10. How does the author define poverty with respect to incomelevel?
11. What examples does he give of patriotism?
Does not include: "Who" when who is an individual rather than collec-tive. Describe is not included unless the question is something like"describe what he means by IiippiJ77-3T unless words other than describeare added to give it additional meaning.
C. Assumptions
Includes questions requiring students to identify assumptionsbeing made which can only be inferred from an analysis of a series ofstatements.
Includes: Questions asking for the identification of an element re-garded as that which is assumed, inferred, implied, presupposed, takenfor granted, underlying, behind, underneath, etc.
Examples
1. What assumptions are being made in this argument?2. What is he inferring about the race riots?3. Is he implying X?1i. What inferences are being made here?5. What is being taken for granted (presupposed) here?6, What must be assumed to underlie that?7. What is behind (underneath) that argument?
149
D. Distinction of Fact from Opinion or Value Judgment
These questions require students to detect the nature andfunction of a particular statement in a communication.
Includes: Questions requiring students to distinguish factual evidencein an argument from opinions or value judgment. (Value judgments referto the good-bad, approve-disapprove, for-against dimension.)
Examples
1. What facts are being presented here?2. What evidence is given in support of the hypothesis?3. How does the author feel about X?Li. What does he think about X?5. What is his idea on this?6. What is his opinion here?7. Is X a fact/ opinion/ value judgment?8. What does he cite as evidence?9. What does he use to defend/back up this hypothesis?
10. What information was that based on?
Does not include: Just because the author said it does not make it afact.
E. Conclusions
Questions requiring students to distinguish a proposed conclu-sion arrived at from supporting statements.
Includes: conclusions, answer, solution, summary point, resolution,final decision arrived at, suggestion made to deal with, alternativesolutions to a problem, ways proposed to deal with a problem, sugges-tions to deal with a problem, etc.
Examples
1. What conclusion does he come to?2. What answer does he arrive at about what the hippies should do?3. Is X the conclusion? (Is that what he is saying? Does he say
that?)
150
4. When he says X, what logical conclusion is he arriving at?5. How does he answer the original question?6. How does he propose to deal with that problem?7. What alternatives to that solution does he offer?
III. QUALITY DIMENSION
A. Identification of Hypotheses - Questions requiring a statement of arelation of dependency among variables.
1. Questions of high quality require students to supply bothvariables involved in the relationship in addition to therelationship itself.
a. What is the hypothesis being irvestigated?b. What is the author trying to prove?c. What relationship is the major premise of his argument?
2. Questions which do not meet the criteria established for highquality include questions in which the student must supply onlyone (or part) of the variables, and/or the relationship itself;
a. What does A cause?b. Why does B happen?c. What is C associated with?d. What is the relationship between A and B?
and questions requiring only that the student agree, disagree,or select from among given alternatives.
a. Is X the hypothesis?b. Is X or Y or Z the hypothesis?
B. Identification of Definitions - Questions requiring students todistinguish the methods and standards used for defining a word.
1. Questions of high quality require students to supply key ele-ments of the definition and their methods of measurement.
a. A proposed civil rights bill would apply to business engag-ing in a "substantial" amount of interstate commerce. Howis "substantial" being defined here?
b. What does the author mean by patriotism?
151
2. Questions which do not meet the criteria established for highquality include questions in which a key element or method ofmeasurement is given;
a. How does the author use the word poverty with respect toincome level?
b. How does the author define patriotism in its military sense?
c. With respect to the amount of tar and nicotine content, howdoes the author define a "revolutionary filter"?
and questions requiring only that the student agree, disagree,or select from among given alternatives.
a. Is X what is meant here by revolutionary?b. With respect to social class, does the author mean income,
education, or occupation?
C. Identification of Assumptions - Questions requiring students tosupply both the assumption and its source (what it concerns or is
in regard to).
1. Questions of high quality require students to supply theassumption and its source.
a. What is implicit in his argument?b. What assumption is the author making here?
2. Questions which do not meet the criteria established for high
quality include, questions in which students are directed to
the source of the assumption;
a. What assumption is he making about the nature of man?
b. What inferences are being drawn about what is of primeimportance to minority groups?
and questions requiring only that the student agree, disagree,
or select from among given alternatives.
a. Is he assuming X or Y about Z?b. Is A the assumption being made?
152
D. Distinction of Factual from Nonfactual Statements - Questionsrequiring students to distinguish the nature and function of aparticular statement in a communication as fact, opinion, or valuejudgment.
1. Questions of high quality require students to supply the fact/opinion/value judgment in its entirety.
a. What evidence is presented in support of the hypothesis?b. What facts are given?c. What value judgments are being made?d. What is the author's opinion of that?
2. Questions which do not meet the criterion established for highquality include questions in which part of the relevant infor-mation is given;
a. What facts are cited about the success of current hearttransplants?
b. What value judgment is being made about the lact of validityin the hippie movement?
and questions requiring only that the student agree, disagree,or select from among given alternatives.
a. Does he think X about Y?b. Is X a fact?c. Is his opinion X or Y?
E. Identification of Conclusions - Questions requiring students todistinguish conclusions from supporting statements?
1. Questions of high quality require students to supply a conclu-sion in its entirety.
a. What conclusion did the author arrive at?b. What solution to the problem is being suggested?c. What alternatives has he suggested to that answer?
2. Questions which do not meet the criterion established for highquality include questions in which part of the informationrelevant to the conclusion;
a. What conclusion did the author come to on what responsibledissenters should do?
153
b. What answer does the author arrive at as to which is thebetter kind of patriotism?
and questions requiring only that the student agree, disagree,or select from among given alternatives.
a. Is the author suggesting that disease prevention is thealternative to heart transplants?
b. Is his solution to the racial problem integration,strengthening of the black community, or both?
APPENDIX B
SET INDUCTION MATERIALS
155
STANFORD UNIVERSITYSchool of Education
Technical Skills Project Prepared by Mary Lou KoranDirector: F. J. McDonald Training Instructions for:
The Development of Analytic Questioning
PLEASE READ THE FOLLOWING INSTRUCTIONS VERY CAREFULLY BEFORE TEACHINGYOUR FIRST LESSON,.
Educators hold that one of the major objectives of instructionin the schools is the development of skills or strategies which will en-able students to become lifelong autonomous learners. While the contentof any field may undergo rapid change, the ability to analyze written ma-terials within that content area is a skill which is always an objectiveof any field of study. One way of developing such ability in students isto train teachers to ask the kinds of questions that require students toengage in analytic thinking with respect to their reading material. Theobjective of this training session is to help you develop the skill ofasking these kinds of questions.
Analytic Questioning
A communication is composed of a large number of elements. Someof these elements are explicitly stated in the communication and can berecognized and classified relatively easily.
There are other elements in a communication which are not soclearly labeled or identified by the writer. These elements can be in-ferred only from an analysis of a series of statements within the communi-cation. Many of these elements may be of great importance in determiningthe nature of the communication. Until the reader can detect them he mayhave difficulty in comprehending and evaluating the communication.
Analytic questioning breaks a communication down into its crit-ical elements or parts such that the function of these elements within thecommunication, as well as the relative hierarchy of ideas, is made clear.Such analyses are intended to clarify the structure of the communication,to indicate how it is organized, and to identify the arrangement of ideasin it.
156
YOUR GOAL IS TO TEACH A LESSON IN SUCH A WAY AS TO ASK QUESTIONS WHICH
LEAD STUDENTS TO IDENTIFY AND CLASSIFY THE CRITICAL ELEMENTS OF THE WRITTEN
COMMUNICATION YOU HAVE BEEN GIVEN.
The written material you have been given is meant to serve as avehicle for analytic questioning. Using the skill of analytic questioningis the objective of the lesson rather than transmitting content.
APPENDIX C
WRITTEN MODEL
157
158
TEACHER: We are dealing with an essay by Archibald McLeish called "When
We Are Gods." You have all had a chance to read this closely now. An
essay, a piece of writing, in fact, any communication can be thought of as
a number of elements together; and usually in a complex interrelationship.So in order to understand the elements that make up a communication, we
need to do some critical analysis. We need to identify these elements,single them out and consider what they are doing in the essay, how they
are contributing to the meaning. So we are going to try to do that with
five categories that relate to the basic thinking process. In fact, this
could be a piece of scientific writing, it could be from a history text
book. As it is, it is an editorial by a philosopher and poet. (NOTICE
THE CATEGORIES OF ANALYTIC QUESTIONS TO BE USED.) The five categories
that we are going to apply to it are: hypotheses, the difference betweenstatements of fact and statements of value, unstated assumptions that he
may be making, definitions and how he is making it known what he means bycertain words, and conclusions that he reaches in the course of his rea-
soning, The categories again are: hypothesis, fact value statements,
basic assumptions, definitions, and conclusions. (NOTICE HOW HE ASKS FOR
HYPOTHESES.) Let's begin with that first category because we could look
for a very broad hypothesis.
STUDENT: I think the idea of the essay is sort of centered around a
paradox that he finds and he brings out in the first paragraph of the rest
of the essay. And the paradox being:, why in this advanced age of ourswith all of our great technological advances, why is man and most of human-
ity still despondent and why does it still feel a great deal of despair
over its own existence.
TEACHER: Okay. He centers then on that paradox. The question seems to
be why the triumph on one side, the uneasiness of humanity on the other.
And notice that he organizes the essay around those two sides: the tri-
umph on the one side, uneasiness on the other. Can you see within the
rest of the essay any place where he seems to be stating another or per-
haps a related hypothesis?
STUDENT: In the third paragraph the first sentence fits nicely into an
"if then" statement. If, in the terms of scientific and industrial accom-plishments, then our age is one of the great ages of history.
TEACHER: Okay. And we can see this as a hypothesis that he goes on to
develop and give some evidence for. (NOTICE HOW HE ASKS FOR THE DISTINC-
TION BETWEEN FACT AND VALUE JUDGMENT.) If we look at it from industrialscientific standpoints, this is one of the great ages of history. Does
he give any facts in relation to this statement? Does he refer to any-
thing that we could call facts rather than value judgmentdor opinions?
159
STUDENT: In the first paragraph, the fact could be the President's state-ment about that nobody seems to see what is right with America, only whatwas wrong with it. A value judgment or a conclusion can be one of thePresident's. "What was right," he said, "was obvious and admirable." Andthen the rest of the essay would probably consist mainly of value judgmentson the part .1f the author.
TEACHER: Okay. Now this is an editorial and I think that is a very impor-tant observation, Debby, that values seem to be throughout it. That he is
making some value judgments. What are some specific value judgments thathe seems to make in the course of the essay?
STUDENT: Well, he makes several through the whole essay. There is one
very blatant one on the second page; it says, "in most ages it is the artswhich are creative and believe; the men of action who despair." And thenhe goes on to say that the arts usually see the truth. And this is apretty extreme assumption, although it may be true.
TEACHER: Okay. The arts see the truth. If that were a statement of factwe would be able to prove it. If it is a statement of value, it would ex-
press something that men believe in. I think we can see that distinction.
(NOTICE HOW HE ASKS FOR DEFINITIONS.) Now, when he gets into talking inthe second half of the essay, notice he is talking about the uneasinessand he uses the term "despair." Does he anywhere in the essay give uswhat we could consider a definition of despair?; What is he talking about
when he says "despair"? Cris?
STUDENT: Uh, I think one probably could say that he says that discoveriesof contemporary literature are old discoveries long since made and thenhe says, the discovery that men each really die and the discovery thatmoral human life is meaningless. He realizes that man is human and thathe realizes that he does possess this technological power, and that he doesrealize that he can destroy himself.
TEACHER: Okay. He has defined despair then as the feeling based on thesediscoveries--old discoveries long since made. Notice him giving examplesthere for definition too because he defines the old discoveries long sincemade. How does he define old discoveries long since made
STUDENT: Well, he gives specific examples of these discoveries.
TEACHER: Right. (NOTICE HOW HE ASKS FOR ASSUMPTIONS.) And that he defines
right within the context as we read what he means. Would you say that overall in his writing that there is any basic unstated assumption, or premisethat he would pretty much hold in mind but not be stating directly;. Think
about that. What do you think? Ken?
160
STUDENT: Possibly the assumption, uh, that the reader is fairly well readin mythology because he brings out several instances, especially near theend, where he brings up Greek Mythology, such as man stealing fire fromthe gods. And also he brings up an analogy of Hercules in the end.
TEACHER: Good. In fact, notice in both cases that he doesn't refer tothe name of the god, the one myth of Prometheus, and he doesn't refer tothe name of Hercules or Heracles. Although he just refers indirectly sowe can conclude that he is assuming some familiarity with myths. Can yousee other basic unstated assumptions within the essay?
STUDENT: Well, kind of, in the next to the last par:4?raph he says, "weknow what we are" and this seems to imply that maybe man has a tendencytowards self destruction, that no matter what happens we'll end up losingourselves.
TEACHER: Good. We know what we are. That almost has a frightening soundto it--we know what we are. (NOTICE HOW HE ASKS FOR CONCLUSIONS.) Whatthen would you say is the conclusion toward which the whole paper movesand would be an answer to the overall problem that he is posing What
is the conclusion?
STUDENT: Well, in the last paragraph i ..is last sentence which would bethe final statement that he makes, he says, like the old Greek here, whowe've already defined as Hercules, who learned when all his labors hadbeen accomplished--that's when all his achievements had been made--thatwould be paralleling Hercules to now, to us, when all our achievementshave been made. That it was he himself who'd killed his sons and that wewould kill our sons, our future with all our achievements--maybe the atombomb or other nuclear weapons.
TEACHER: Good. And that bringing in the myth at the end, you see, espe-cially to someone familiar with the myth, would raise a kind of frighten-ing idea: The old Greek completing his labors;iHercules completing hislabors just like science completes its labors acid then going ins:*,ne andkilling the future, killing his sons. Overall then, we see McLeish rea-soning through, giving in many cases poetic examples. And let's see ifwe can summarize very quickly and see for instance how all that we've said
in a way relates to the hypotheses. What was the major hypothesis that he
dealt with?
STUDENT: Why is there this paradox of great achievement and at the same
time despair.
TEACHER: Right. Why the. . .why the achievement on one side and yet the
almost total despair. And notice how each part of the essay relates to
161
trying to answer that question. Either to justify his statement thatthere has been great triumph in technological scientific areas, or to showthe despair--and then toward the end to answer "Why." And basically, thatanswer, why the despair, is what?
STUDENT: Uhl man is afraid of self-destruction. .
TEACHER: Right.
STUDENT: . . .afraid of his power.
TEACHER: Man is afraid of self-destruction, afraid of his power. .
STUDENT: We know what we are. .
TEACHER: We know what we are. A good sentence to close with--we know whatwe are.
APPENDIX D
WRITTEN MEASURES
163
Below is a list of kinds of questions possible for classroom use. Place
a "T" or "F" in the space provided at the left of each item number
according to whether or not it represents a major kind or category of
Analytic Questioning.
1. Identification of observations
2. Identification of experimental procedures
3. Identification of hypotheses
4. Identification of inferences
5. Identification of theories
6. Identification of conclusions
7. Identification of unstated assumptions
8. Identification of predictions
9. Identification of factual vs. nonfactual statements
10. Identification of manipulation or control of variables
11. Identification of generalizations
12. Identification of semantic definitions
13. Identification of intent of author
14. Identification of the classification of objects or observations1111M110
164
Below is a list of the five major categories of Analytic Questioningfollowed by a list of questions. Identify which of the categories eachquestion fits into by writing the number of the category in the space
provided at the left of the sentence. If a question does not correspond
to any of the given categories, leave the space blank.
1. Identification of hypotheses
2. Identification of factual vs. nonfactual statements
3. Identification of assumptions4. Identification of semantic definition
5. Identification of conclusions
What does the author believe has caused an increase in the use of
narcotics?
Into which categories would Hemingway's writing fall?
What is the writer's opinion on that question?
A proposal for a civil rights bill would apply to businesses en-
gaging in a substantial amount of interstate commerce. What is
meant here by substantial?
Which statement is the logical outcome of his argument?
How could you predict the form of the written material if you knew
its function?
Could you state something that you believe is true for all of these
cases?
What statement constitutes the major premise of the argument?
What would happen if I increased the temperature of the water?
What does the author think is the relationship between poverty and
crime?
What is the answer to that question in your own opinion?
What kind of an experiment could you set up to test the theory
that form determines function?
What inference could you make from these observations?
How does the writer feel about student demonstrations?
What underlying belief is implicit in this argument?
What information would we need to have about that?
What can be inferred in the data from which Galileo extrapolated
the case for free fall?
What examples are given here of "poverty and deprivation"?
165
APPENDIX
RATING FORMS
166
RATING FORMS
TOTAL ANALYTIC QUESTIONS
Hypotheses
Definitions
Assumptions
Fact - Nonfact
Conclusions
TOTAL CATEGORIES
TOTAL HIGH LEVEL
TOTAL NONANALYTIC QUESTIONS
NAME:
DATE:
SCRIPT:
167
NAME:
DATE:
SCRIPT:
ANALYTIC QUESTIONSTotal
HYPOTHESES:i
High Quality
Low Quality
DEFINITIONS:
High Quality
Low Quality
ASSUMPTIONS:
High Quality
Low Quality
FACT - NONFACT:
High Quality
Low Quality
CONCLUSIONS:
High Quality
Low
NONANALYTIC QUESTIONSTotal
APPENDIX F
INTERCORRELATIONS AMONG DEPENDENT VARIABLES
169
INTERCORRELATIONSAMWDEPENDENT VARIABLESFILM-MEDIATED MODELING GROUP
2 3 4 5 6 7 8 9 10 11 12 13 14
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Analytic Questions, T1
Categories; T1.
High Quality, T1
Nonanalytic, T1
Analytic Questions, T2
Categories, T2
High Quality, T2
Nonanalytic, T2
Analytic Questions
Categories, T3
High Quality, T3
Nonanalytic, T3
True-False Test
Matching Test
80 83
72
17
15
15
33
11
26
13
20
11
06
-20
53
33
17
29
11
92
54
00
11
-10
43
14
-05
07
13
01
16
08
52
27
44
-04
07
01
09
-17
50
49
50
-08
68
17
04
27
11
50
2
48
-18
91
66
-05
03
-11
54
02
-26
-03
64
-16
-16
-23
07
-07
00
08
22
-08
12
-07
24
07
26
07
16
19
07
06
26
30
20
11
25
12
25
-05
33
Decimals omitted
INTERCORRELATIONS AMONG DEPENDENT VARIABLESWRITTEN MODELING GROUP
2 3 4 5 6 7 8 9 10 11 12 13 14
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Analytic Questions, T1
Categories, T1
High Quality, T1
Nonanalytic, T1
Analytic Questions, T2
Categories, T2
High Quality, T2
Nonanalytic, T2
Analytic Questions, T3
Categories, T3
High Quality, T3
Nonanalytic, T3
True-False Test
Matching Test
78 87
70
12
21
17
48
52
42
29
32
34
29
23
77
54
51
49
20
94
73
-20
-18
-13
00
40
-42
-42
50
41
37
35
69
56
60
-30
38
32
19
15
52
58
44
-23
74
47
37
39
29
67
58
62
-26
91
73
-09
-02
-03
25
24
-33
-29
58
-19
-27
23
14
23
00
18
48
47
46
-22
46
48
46
-07
-37
-17
-40
-11
-04
10
-06
-04
-06
17
-06
-04
49
Decimals omitted
INTERCORRELATIONS AMONG DEPENDENT VARIABLESCONTROL GROUP
2 3 4 5 6 7 8 9 7_0 11 12 13 14
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Analytic Questions, T1
Categories: T1
High Quality, T1
Nonanalytic, T1
Analytic Questions, T2
Categories, T2
High Quality, T2
Nonanalytic, T2
Analytic Questions, T3
Categories, T3
High Quality, T3
Nonanalytic, T3
True-False Test
Matching Test
72 79
61
12
17
18
46
39
27
-17
42
40
31
-12
79
43
35
37
-09
73
66
12
15
02
58
-01
00
-07
70
47
68
12
47
47
46
15
55
39
44
02
53
60
47
07
79
59
42
68
17
28
32
38
17
91
68
37
34
32
49
18
35
25
58
45
43
38
08
-04
-09
16
00
-12
00
18
16
-10
-19
-05
00
-05
-11
-29
03
-18
-12
-03
-03
-01
00
-07
.24
Decimals omitted
APPENDIX G
ANALYSES FOR HOMOGENEITY ASSUMPTION
173
SUMMARY OF REGRESSION ANALYSES BETWEEN TEACHING SESSIONS
Variables
Treatment Group
Film-Mediated
Modeling
ab
Written
Modeling
ab
Control
a
Analytic Questions, T1 -T2
11.10
.64
4.89
.90
1.84
.40
1.05
NS
Analytic Questions, T2 -T3
6.59
.52
2.88
.88
.99
.59
2.14
NS
Analytic Questions, T1 -T3
12.70
.25
5.79
1.18
.17
.74
3.15
.05
Categories of Analytic Questions, T1 -T2
3.55
.12
2.07
.47
.90
.39
1.13
NS
Categories of Analytic Questions, T2 -T3
2.06
.48
1.32
.57
.49
.53
.16
NS
Categories of Analytic Questions, T 1-T 3
3.83
.01
2.22
.44
.65
.39
2.02
NS
High Quality Analytic Questions, T1 -T2
8.49
.74
3.11
1.00
1.19
.33
1.46
NS
High Quality Analytic Questions, T2 -T3
5.09
.51
2.98
.72
.70
.51
.64
NS
High Quality Analytic Questions, T1 -T3
8.73
.76
4.83
.91
.19
.68
.14
NS
Nonanalytic Questions, T1 -T2
6.04
.43
19.42
-.01
11.0
.57
3.64
.05
Nonanalytic Questions, T2- T3
6.47
.56
8.36
.46
7.74
..72
1.12
NS
Nonanalytic Questions, T1 -T3
4.50
.47
11.07
.25
10.07
..62
1.60
NS
Note:
a and b take the form:
y = a + bx
-4
VARIANCE-COVARIANCE MATRICES
Control Written Modeling Film-Mediated Modeling
ANALYTIC QUESTIONS
T1 T2 T3 T1 T2 T3 T1
T2 T3
T1 9.40 3.35 7.15 9.02 8.17 10.73 11.37 7.25 2.79
T2 3.35 5.65 3.68 8.17 31.28 22.63 7.25 41.06 21.42
T3 7.15 3.68 10.85 10.73 27.63 50.61 2.79 21.42 40.08
CATEGORIES OF ANALYTIC QUESTIONS
T1 T2 T3 T1 T2 T3 T1 T2
T3
T, 1.08 .44 .40 1.03 .48 .45 1.27 .16 .01
T2 .44 1.08 .61 .48 1.93 1.11 .16 1.42 .68
T3 .40 .61 .94 .45 1.11 1.89 .01 .68 1.36
HIGH QUALITY ANALYTIC QUESTIONS
T1 T2 T3 T1 T2 T3 T1 T2 T3
T1 4.96 1.33 3.46 3.60 3.60 3.31 4.11 3.06 3.11
T2 1.33 2.50 1.38 3.60 14.78 10.68 3.06 27.05 13.83
T3 3.46 1.38 5.20 3.31 10.68 19.60 3.11 13.83 30.54
NONANALYTIC QUESTIONS
T1 T3
T T2 T3
T11
T2 1
T T2 T3
Tl 114.89 67.17 68.45 87.16 -1.06 22.04 88.66 38.04 41.94
T2 67.17 116.39 81.78 -1.06 132.05 61.27 38.04 87.16 49.45
T3 68.45 81.78 165.37 22.04 61.27 83.84 41.94 49.45 66.70
APPENDIX H
ANALYSES OF VARIANCE
CLASSROOM PERFORMANCE MEASURES
176
ANALYSIS OF VARIANCE OF ANALYTIC QUESTIONS
Source Variation df MS
177
Teaching Session 1
Between Groups 2 .16
Within Groups 118 10.01
.02 NS
Teaching Session 2
Between Groups
Within Groups
2 1030.26 39.13 .01
118 26.32
Teaching Session 3
Between Groups 2 1191.01 35.25 .01
Within Groups 118 33.80
178
ANALYSIS OF VARIANCE OF CATEGORIES
OF ANALYTIC QUESTIONS
Source of Variation df MS
Teaching Session 1
Between Groups 2 1.34 1.13 NS
Within Groups 118 1.18
Teaching Session 2
Between Groups 2 46.91 31.47 .01
Within Groups 118 1.49
Teaching Session 3
Between Groups 2
Within Groups 118
66.05 47.13 .01
1.40
179
ANALYSIS OF VARIANCE ON HIGH QUALITY
ANALYTIC QUESTIONS
Source of Variation
Between Groups
Within Groups
Between Groups
Within Groups
Between Groups
Within Groups
df MS
Teaching Session 1
2 .90
118 4.25
.22 NS
Teaching Session 2
2 652.46
118 15.01
45.46 .01
Teaching Session 3
2 734.12
118 18.49
39.69 .01
ANALYSIS OF VARIANCE OF NONALYTIC QUESTIONS
180
Source of Variation df MS F p
Teaching Session 1
Between Groups 2 39.73
Within Groups 118 97.45
.41 NS
Teaching Session 2
Between Groups 2 92111 8.35 .01
Within Groups 118 110.34
Teaching Session 3
Between Groups 2 1262.89 12,17
Within Groups 118 103.80
.01
181
ANALYSIS OF VARIANCE OF ANALYTIC QUESTIONS
Source df MS F
Film-Mediated Modeling Group
TeachingPersonsResidual
Sessions 2 1335.5139 51.8278 20.35
65.62
Written Modeling Group
TeachingPersonsResidual
Sessions 2 442.0039 61.3378 14.79
29.86
Control Group
TeachingPersonsResidual
Sessions 2 6.3039 18.5078 3.93
1.60
p
.01
.01
NS
182
ANALYSIS OF VARIANCE OF CATEGORIES OF ANALYTIC QUESTIONS
Source df MS
Film-Mediated Modeling Group
TeachingPersonsResidual
Sessions 2 75.6539 1.9278 1.06
70.70 .01
Written Modeling Group
TeachingPersonsResidual
Sessions 2 18.9839 2.9978 .93
20.40 .01
Contol Group
TeachingPersonsResidual
Sessions 2 1.8339 2.1378 .58
3.15 NS
183
ANALYSIS OF VARIANCE OF HIGH
QUALITY ANALYTIC QUESTIONS
Source df MS
Film-Mediated Modeling Group
TeachingPersonsResidual
Sessions 2 906.2739 33.9178 13.89
65.19 .01
Written Modeling Group
TeachingPersonsResidual
Sessions 2 225.5239 24.4078 6.79
33.16 .01
Control Group
TeachingPersonsResidual
Sessions 2 2.5039 8.7278 2.11
1.18 NS
184
ANALYSIS OF VARIANCE OF NONANALYTIC QUESTIONS
Source df MS
Film-Mediated Modeling Group
TeachingPersonsResidual
Sessions 2
3978
772.12167.3037.86
20.10 .01
Written Modeling Group
TeachingPersonsResidual
Sessions 23978
550.22155.8573.60
7.47 .01
Control Group
TeachingPersonsResidual
Sessions 2
3978
5.602749557.73
.10 NS
APPENDIX I
TESTS ON MEANS USING NEWMAN-KEULS PROCEDURE
CLASSROOM PERFORMANCE MEASURES
185
ANALYTIC QUESTIONS - TEACHING SESSION 2
186
TreatmentGroups
111.1111. Control 2. Written 3. Film
Mediated
Ordered Means 3.32 8.16 13.40
Differences be-tween pairs
1.
2.
4.84** 10.08**
5.24**
SA= .77 r
2r3
q .95 2.80 3.36
q .99 3.70 4.20
SA-cl 952.15 2.58
S-107.q .99 2.85 3.23
SOMIIIMM=141MIIINIMEMMEMOMMAIMMIM=Y1
*p < .05
**p < .01
TreatmentGroups
ANALYTIC QUESTIONS - TEACHING SESSION 3
.IIaMIMmMMIMIMII=MMtMms1. Control
Ordered Means 2.93
Differences be-tween pairs
1.
187
2. Written 3. FilmMediated
10.09 13.58
7.16** 10.65**
3.49**
SA= .77 r
2r3
q .95 2.80 3.36
q .99 3.70 4.20
SAq .95 2.15 2.58
SA q .99 2.85 3.23
*p < .05
**p < .01
188
CATEGORIES OF
ANALYTIC QUESTIONS - TELCHING SESSION 2
IMIIMMIMIIMMI=1.1 .1111.11111=01=PMWMIIIMMIM
Treatment 1. Control 2. Written 3. FilmGroups Mediated
Ordered Means 1.59 2.89
Differences be-tween pairs
1.
2.
1.30**
3.73
2.14**
.84**
SA
= .18 r2
r3
q .95 2.80 3.36
q .99 3.70 4.20
Szi .95 .40 .60A
SA q .99 .66 .75
.111WMOOMOOM401...........1.. .=.0*p < .05
**p < .01
189
CATEGORIES OF
ANALYTIC QUESTIONS - TEACHING SESSION 3
TreatmentGroups
.11111.111M.IND ONO
1. Control 2. Written 3. FilmMediated
Ordered Means 1.33 2.98
Differences be-tween pairs
1.
2.
1.65**
3.84
2.51**
.86**
SA= .18 r
2 3
q .95 2.80 3.36
q .99 3.70 4.20
Sq .95 .40 .60A
SA
.99 .66 .75.*p < .05
**p < .01
IIIINNIIINII=
190
HIGH QUALITY
ANALYTIC QUESTIONS - TEACHING SESSION 2
TreatmentGroups
1. Control
im1M.M.I.W.N111m...W11.14.11MiMIIMASOM.1.11111IIIMI.
2. Written 3. FilmMediated
Ordered Means 1.92 5.18 9.90
Differences be-tween pairs
1.
2.
3.26** 7.98**
4.72**
SI= .57 2r3
q .95 2.80 3.36
q .99 3.70 4.20
SA 4 .95 1.60 1.92
Sq .99 2.11 2.39A
*p < .05
(.1) < .01
191
HIGH QUALITY
ANALYTIC QUESTIONS - TEACHING SESSION 3
Treatment 1. Control 2. Written 3. FilmGroups Mediated
Ord d Means 1.68 6.73 10.15
Differences be-tween pairs
1. 5.05** 8.47**
2. 3.42**
S = .57A
r2 r3
q .95 2.80 3.36
q .99 3.70 4.20
.95 1.60 1.92A
.99 2.11 2.39A
*p < .05
**p < .01
{.
NONANALYTIC QUESTIONS - TEACHING SESSION
192
2
TreatmentGroups
1. FilmMediated
2. Written 3. Control
Ordered Means 16.02 19.13 25.40
1.Differences be-tween pairs
2.
3.11 9.38**
6.27**
SA= 1.61 r
2r3
q .95 2.80 3.36
q .99 3.70 4.20
SAq .95 4.51 5.40
SAq .99 5.96 6.76
*p < .05
**p < .01
"T.
193
NONANALYTIC QUESTIONS - TEACHING SESSION 3
TreatmentGroups
1. Film 2.
MediatedWritten 3. Control
Ordered Means 15.51 16.24 25.96
Differences be-tween pairs
1.
2.
1.73 10.45**
8.72**
SA = 1.61 r2 r3
q .95 2.80 3.36
q .99 3.70 4.20
SA q .95 4.51 5.40
SA q .99 5.96 6.76
*p < .05
**p < .01
APPENDIX J
TESTS ON MEANS USING NEWMAN-KEULS
PROCEDURE WRITTEN MEASURES
194
MATCHING TEST
195
TreatmentGroups
1. Control 2. Written 3. FilmMediated
Ordered Means 7.03 9.77 10.71
Differences be-tween pairs
1.
2.
2.74** 3.68**
.94*
11111101110 .33 r2 r
3
q .95 2.80 3.36
q .99 3.70 4.20
Sq .95 .92 1.10A
S-41 .99 1.22 1.40
*p < .05
**p < .01
TreatmentGroups
196
TRUE-FALSE TEST
1. Control 2. Written 3. FilmMediated
Ordered Means 7.90 10.67 11.83
Differences be-tween pairs
1. 2.77** 3.93**
2. 1.16
ellne011111S.44
A 2 3
q .95 2.80 3.36
q .99 3.70 4.20
Sxq .95 1.23 1.48
STiq .99 1.60 1.85
*p < .05
**p < .01
APPENDIX K
INTERCORRELATIONS OF ABILITY MEASURES
197
INTERCORRELATIONS OF ABILITY MEASURES:
FILM-MEDIATED MODELING GROUP
12
34
56
78
910
11
12
13
14
15
16
17
18
19
20
21
22
23
1.
Hidden Figures - Part I
59
90
43
32
39
40
43
45
32
08
18
03
13
10
31
-13
25
10
19
07
15
13
2.
Hidden Figures - Part II
87
31
30
33
29
38
37
22
09
16
1P
23
22
24
Ott
19
02
13
07
30
22
3.
Hidden Figures - Total
42
35
40
39
46
46
31
09
20
11
20
18
31
-05
25
08
18
08
25
20
4.
Identical Pictures - Part I
65
83
41
49
48
24
13
20
32
33
34
32
-05
22
37
34
07
14
12
5.
Identical Pictures - Part II
93
40
48
48
09
18
15
21
34
29
25
00
07
23
17
-02
09
04
6.
Identical Pictures - Total
46
54
54
20
19
22
29
37
34
32
-01
12
26
22
-01
09
05
7.
Maze Tracing - Part I
72
91
18
02
12
25
25
26
40
09
14
17
21
CO -03
-02
8.
Maze Tracing - Part II
93
36
26
35
37
25
33
36
00
-01
17
10
-15 -13
-15
9.
Maze Tracing - Total
30
16
27
34
27
32
41
04
05
18
16
-09 -09
-10
10.
Word Arrangement - Part I
68
88
46
29
39
33
10
24
29
28
08
15
14
11.
Word Arrangement - Part II
93
53
40
48
36
00
09
27
22
-03
12
06
12.
Word Arrangement - Total
56
39
49
39
04
17
30
28
04
14
11
13.
Advanced Vocabulary - Part I
82
94
67
00
21
28
32
23
23
25
14.
Advanced Vocabulary - Part II
95
69
00
30
26
35
33
38
39
15.
Advanced Vocabulary - Total
72
08
27
28
35
29
32
34
16.
Verbal GRE
00
26
22
32
31
25
29
17.
Film Memory
28
41
35
13
11
13
18.
Memory for Ideas - Part I
69
91
59
67
69
19.
Memory for Ideas - Part II
89
40 44
46
20.
Memory for Ideas - Total
55
60
63
21.
Sentence Reproduction - Part
I68
88
22.
Sentence Reproduction - Part
II
94
23.
Sentence Reproduction - Total
Decimals omitted.
INTERCORRELATIONS OF ABILITY<I4EASURES:
WRITTEN MODELING GROUP
12
34
56
78
910
11
12
13
14
15
16
17
18
19
20
21
22
23
1.
Hidden Figures - Part I
50
86
18
36
29
26
32
Z1
25
14
20
19
24
22
17
13
02
-04
01
-19
02-14
2.
Hidden Figures - Part II
87
27
48
42
14
20
19
26
15
24
20
14
19
08
32
25
13
25
-06
-01
-09
3.
Hidden Figures - Total
26
49
41
23
30
29
29
17
25
23
22
23
14
26
16
05
15
-14
00
-13
4.
Identical Pictures - Part I
75
90
33
31
34
26
25
25
35
26
33
19-05
37
21
35
31
21
26
5.
Identical Pictures - Part II
93
41
42
45
21
27
24
37
32
37
27
11
29
15
26
22
12
15
6.
Identical Pictures - Total
43
41
45
26
29
28
42
31
39
30
02
37
18
33
26
17
21
7.
Maze Tracing - Part I
72
91
10
32
22
11
28
23
31
-03
02
24
67
17
25
21
8.
Maze Tracing - Part II
94
09
23
19
26
34
33
22
01
-04
08
-01
23
26
24
9.
Maze Tracing - Total
10
29
22
20
34
31
28
02
-01
16
03
21
28
25
10.
Word Arrangement - Part I
73
90
-07
-01
-05
06
18
15
11
19
13
02
05
11.
Word Arrangement - Part II
94-02
07
01
07
04
18
06
-16
16
10
10
12.
Word Arrangement - Total
-06
03
-02
06
11
15
06
15
15
07
09
13.
Advanced Vocabulary - Part I
77
93
39-15
00
03
-03
25
-02
09
14.
Advanced Vocabulary - Part II
93
42
-01
-03
06
-04
32
13
22
15.
Advanced Vocabulary - Total
43
-08
00
06
-02
29
08
18
16.
Verbal GRE
-06
01
03
00
25
31
36
17.
Film Memory
17
20
30
-13
-12
-16
18.
Memory for Ideas - Part I
32
92
20
12
16
19.
Memory for Ideas - Part II
59
22
-03
07
20.
Memory for Ideas - Total
23
07
14
21.
Sentence Reproduction - Part
I43
84
22.
Sentence Reproduction - Part
II
81
23.
Sentence Reproduction - Total
Decimals omitted.
INTERCORRELATIONS OF ABILITY MEASURES:
CONTROL GROUP
12
34
56
78
910
11
12
13
14
15
16
17
18
16
20
21
22
23
1.
Hidden Figures - Part I
73
93
31
30
33
27
09
19
18
05
11
-02
08
03
00-15
-05
-20
-13
13
-05
02
2.
Hidden Figures - Part II
92
41
51
49
30
28
32
24
15
18
-13
02
-04
-02
-05
-04
00
-02
-01
-09
-07
3.
Hidden Figures - Total
39
43
44
31
19
27
22
11
16
-08
05
00
00
-11
-05
-11
-08
06
-07
-02
4.
Identical Pictures - Part I
74
92
40
49
48
12
15
15
11
34
26
37
-02
00
03
00
15
27
26
5.
Identical Pictures - Part II
92
48
48
51
-07
11
06
17
33
28
27
07
-05
10
01
11
22
21
6.
Identical Pictures - Total
47
52
53
-01
15
06
15
38
30
34
05
00
10
04
14
29
27
7.
Maze Tracing - Part I
80
94-05
-02
-03
-10
11
01
29
42
25
16
25
10
08
11
8.
Maze Tracing - Part II
94
01
11
08
-19
02
-07
20
37
25
16
25
04
03
04
9.
Maze Tracing - Total
-02
04
01
-15
06
-03
26
41
27
17
26
06
05
07
10.
Word Arrangement - Part I
66
89-04
00-02
08
12
37
01
25
00
02
01
11.
Word Arrangement - Part II
91
01
06
04
14
14
41
23
39
-07
00
-03
12.
Word Arrangement - Total
-01
06
03
16
14
43
14
35
-01
03
01
13.
Advanced Vocabulary - Part I
75
92
50
13
-01
-01
-01
25
54
50
14.
Advanced Vocabulary - Part II
95
68
25
00
-01
00
32
46
47
15.
Advanced Vocabulary - Total
64
21
-01
-01
-01
31
53
52
16.
Verbal GRE
23
02
12
07
27
41
41
17.
Film Memory
46
44
52
08
17
16
18.
Memory for Ideas - Part I
50
91
-07
14
06
19.
Memory for Ideas - Part II
81
02
00
01
20.
Memory for Ideas - Total
-03
10
05
21.
Sentence Reproduction - Part
I42
77
22.
Sentence Reproduction - Part
II
90
23.
Sentence Reproduction - Total
Decimals omitted.
O O
BIBLIOGRAPHY
.
a
201
z.
BIBLIOGRAPHY
Allison, R. B., Jr. Learning parameters and human abili-ties, Princeton, N.J.: Educational TestingService, 1960.
Atkinson, J. W. The psychology of motivation. Princeton:Van Nostrand Co., Inc., 1964.
Bandura, A. Behavior modification through modeling proce-dures. In Krasner and Ullman (Eds.) Researchin behavior modification. New York: Holt, Rine-hart and Winston, 1964.
Bandura, A. Social learning through imitation. Nebraskasymposium on motivation. In. M. K. Jones, ed.,Lincoln: U. of Nebraska Press, 1962, 211-269.
Bandura, A..; and Mischel, W. Modification of self imposeddelay of reward through exposure to live andsymbolic models. J. of pers and soc Psychol,1965, 2:698-705.
Bandura, A., Ross, D., and Ross, S. Imitation of silmmediated aggressive models. J. of ab. and soc.Psychol., 1963, 6, 3-11.
Bandura, A. and Walters, R. Social learning and person-ality development, Chicago: Holt, Rinehart andWinston, 1963.
Blaine, D. D., and Dunham, J. L. Memory abilities in con-cept learning. Paper presented at the meetingof the Southwestern Psychological Association,New Orleans, April, 1968.
Bloom, B. S. (Ed), Taxonomy of educational objectives,New York: Longmans, Green, 1956.
Botzum, W. A. A factorial study of the reasoning andclosure factors. Psychometrika, 1963, 16, 361-386.
Bunderson, C. V. The computer and instructional design.Paper presented at the Conference on ComputerAssisted Instruction, University of Texas, Octo-ber, 1968.
202
203
Carroll, J. B. Study of language. Cambridge, Mass.:Harvard University Press, 1959.
Cronbach, L. J. The two disciplines of scientific psy-chology. American Psychologist, 1957, 12, 671-684.
Cronbach, L. J. Essentials of psychological testing. NewYork: Harper & Row, 1963.
Cronbach, L. J. How can instruction be adapted to indi-vidual differences. In Gagne, R., Learning andindividual differences. Columbus, Ohio: CharlesE. Merril Books Inc., 1967.
Cronbach, L. J. "Individual differences in learningability as a function of instructional variables,"Basic and Applied Research Proposal, Submittedto the U.S. Commission of Education, August 15,1965.
Dixon, W. J. (Ed.), Biomedical computer programs. LosAngeles: School of Medicine, University ofCalifornia, Los Angeles, Revised September,1965.
Duncanson, J. P. Intelligence and the ability to learn.Princeton, N.J.: Educational Testing Service,1964.
Edgerton, H. E. The interaction of trainee aptitude pat-tern with method of instruction. New York:Richardson, Bellows, Henry and Co., 1958.
Fitts, P. M. Perceptual experiments and process theory.In Gagnd, R., Learning and individual differ-ences. Columbus, Ohio; Charles E. Merril BooksInc., 1967.
Fleishman, E. A. The description and prediction of per-ceptual-motor skill learning. In R. Glaser(ed.) Training research and education. NewYork: John Wiley & Sons, Inc., 1965.
204
French, fJ Ekstrom, R., and Price, L. L Manual for kit ofreference tests for cognitive factors. (Revised1963), Princeton, New Jersey: Educational Test-ing Service, 1963.
Gagn41 R. Problem solving in A. W. Melton (Ed.) Cate-zories of human learning, New York: AcademicPress, 1964, p. 293-317.
Gagn4, R. M., and Paradise, N. E Abilities and learningsets in knowledge acquisition. Psychol. Monogr.1961, 75, No. 14 (Whole No. 518).
Glaser, R. Some implications of previous work on learningand individual differences. In Gagne, R., Learn-ing and individual differences. Columbus, Ohio:Charles E. Merril Books, Inc., 1967.
Grimes, J. W. and Allinsmith, W. Compulsivity, anxietyand school achievement. Merril Palmer Quarterly,1961, 7, 247-271.
Humphreys, L. G. Some investigations of the simplex.Psychometrika, 1960, 25, 313-323.
Husband, R. W. Intercorrelations among learning abilities:I, J. genet. Psychol., 1939, 55, 353-364.
Husband, R. W. Intercorrelations among learning abilities:III, The effect of length of tests. J. ofgenet. Psychol., 1941, 58, 427-430.
Johnson, P. 0., and Neyman, J. Tests of certain linearhypotheses and their application to some educa-tional problems. Statis. res Mem, U. of London,1936, 1, 57-93.
Koran, J. J. The relative effects of imitation versusproblem solving on the acquisition of a complexteaching skill, an unpublished dissertation,Stanford University, 1968.
Lloyd Morgan, C. Habit and Instinct. London: Arnold, 1896.
Manning, W. H., and Dubois, P. H. Correlational methodsin research on human learning. Perceptual andMotor Skills, 1962, 15, 287-321, MonographSupplement 3-V15.
205
McBrearty, J. F., Marston, A. F., and Kanfer, F. H. Con-ditioning a verbal operant in a group setting:direct versus vicarious reinforcement. AmericanPsychologist, 1961, 16, 425-434.
McDonald, F. J. and Allen, D. W. Training effects offeedback and modeling procedures on teachingperformance. Stanford University, 1967.
McGuire, W. J. A multiprocess model for paired-associate learning. J. Exp. Psychol., 1961,62:335-347.
Melton, A. W. Individual differences and theoreticalprocess variables: General comments on theconference. In Gagn4, R. (ed.) Learning andIndividual Differences. Columbus, Ohio: CharlesE. Merril Books, Inc., 1967.
Miller, N. E. and Dollard, J. Social learning and imitation.New Haven: Yale University Press, 1941.
Mowrer, O. H. Learning theory and the symbolic processes.New York: Wiley, 1960.
Pemberton, C. L. The closure factors related to othercognitive processes. Psychometrika, 17, 267-288.
Porter, D. Some effects of year long teaching machineinstruction, in E. Galanter (Ed.) AutomaticTeaching. New York: Wiley.
Rapier, J. L. Measured intellifrence and the ability tolearn. Acta Psychologica, 1962, 20, 1-17.
Sanders, N. Classroom questions. New York: Harper andRowe, 1966.
Seibert, W. F., and Snow, R. E. Studies in cine-psychometryI. Final Report, July, 1965, USOE, Grant Number7-12-0280-184.
Seibert, W. F., Reid, C., and Snow, R. Studies in cine-psychometry II. Final Report, December, 1967,USOE, Grant Number 7-24-0880-257.
206
Sheffield, F. D. Theoretical considerations in the learn-ing of complex sequential tasks from demonstra-tion and practice. In A. A. Lumsdaine (Ed.).Student res onse in ro rammed instruction: Asymposium. Wastangton D.C.: National Academyof Sciences--National Research Council, 1961,pp. 13-32.
Simrall, D. Intelligence and the ability to learn. Journalof Psychology, 1947, 28, 27-43.
Snow, R. E., Tiffin, J., Siebert, W. Individual differ-ences and instructional film effects, J. of ed.Psychol., 1965, 56, 315-326.
Stake, R. E. Learning parameters, aptitudes and achieve-ment. Psychometric Monographs. PsychometricSociety, 1961, No. 9.
Tarde, G. The laws of imitation. New York: Holt, 1903.
Thurstone, L. L. A factor analysis study of perception.Psychometr. Monogr., 1944, No. 4.
Wallach, M., and Kogan, N. Risk Taking. New York: Holt,Rinehart & Winston, Inc., 1964.
White, B. W. Visual and auditory closure, J. exp. Psychol.,1954, 48, 234-243.
Winer, B. J. Statistical principles in experimental de-sign. New York: McGraw-Hill, 1962.
Witkin, H. A., Dyk, Ruth B., Faterson, Hanna F., Goodenough,D. R., and Kark, S. A. psychological differen-tiation: Studies of development. New York:Wiley, 1962:
Woodrow, H. The relation between abilities and improve-ment with practice. J. of educ. Psychol., 1938,29, 215-238.
Woodrow, H. Interrelation of measures of learning. J.Psychol., 1940, 10, 49-73.
A.
Woodrow, H. Factors ln improvement with practice. J.Psychol., 1939, 7, 55-70.
Woodrow, H. The ability to learn. Psych. Rev., 1946,53, 147-158.
207
kaArk Anep typing service
2707 HEMPHILL PARK AUSTIN, TEXAS 78705 AC 512 472-3210